Keith's Web Blog RSS Feed

Fedora Core 2 Beta

Sat, 14 Feb 2004 12:04:42 -0700

Looking forward to using Fedora Core 2. Got the beta version finally but haven't installed it onto any system. Will use my laptop as a test system. I wonder if my Sony DVD player will be recognize with this? Somethings I'd like to see: 1) General performance; How does the linux 2.6.1 kernel perform? 2) KDE 3.2; they just released the stable version but the one on Core 2 beta is "just" 3.1.95 (or that's what the guys at distrowatch.com say) 3) KOffice 1.3. I mostly use Open Office, but I wouldn't mind giving this one a shot. Besides those, I don't really see many library changes that I'm not already using in place. I heard the new version of Gnome will be included, but I'm not really a Gnome user. Hopefully things look good though. Somethings I wish they would include: 1) JRE/JDK; it's a major pain getting an updated version of java installed on the system everytime I do an install. 2) Tomcat/JBoss. There might be licensing issues with JBoss, but Tomcat should be an rpm that's integrated. I'd like to see a fully integrated suite of Apache + mod_ssl + mod_perl + tomcat + mod_python + php. It looks like crap (because of potential resource consumption) but there should be enough variations and builds in the RPM package that allows for any number of combination of these packages based on business needs. I tend to build my own web server suite rather than use the distributions. It kinda sucks because it increases my overhead time that I can use spent on development. Make note Red Hat! 3) More intuitive configuration tools. This is where Red Hat could really benefit the open source community and vice versa. The other day I had to deal with sendmail. Anyone who's worked with this load knows what a nightmare it is. Fortunately, I had another server I was migrating off of so I had an example. But imagine a newbie who wants a mail server up and running. Sure in the long term you want to learn how it really works, but opening up something like sendmail is like going to McDonald's factories and seeing how they really manufacture their food. Hardcore linux people will hate me for saying this, but there needs to be more configuration GUI tools that are intuitive. Not all of us enjoy spending all day and all night pouring through manuals and online documentation just to get their mail server working. The backend should be left open for experts, but in general there really needs to be more easy to use tools for these purposes. 4) Integrated apt-get/synaptic tool. I see these tools as the future of Red Hat. The base RPM system needs to be revised, imo, because it's too slow and cumbersome. Maybe an XML tree. But there's been more than one time where my system gets screwed up because of a bad RPM install. One major problem is dependencies between packages. It's always been a problem. apt-get solves a lot of these similar to how Debian works. Synaptic is the thing that will speak to the average user. One problem with Synaptic though is speed and occasional breakdowns. I've seen this tool just freeze for god knows what reason. It has to improve on stability and speed. My theory is to have it integrate with something like BitTorrent when grabbing packages to increase network speed and then provide a better GUI. The idea is good. But there's tons of improvements that can be made so the public can accept this.

PHP

Sun, 15 Feb 2004 12:33:10 -0700

I was pretty resistant to PHP for the longest time. Now, I'm obviously using it and find it to be quite acceptable. My voyage into this web programming language was simple: I'm bored with Perl. Been doing perl too long (5 years) and feel I can do enough where I don't need to push myself further. Instead, I wanted to explore more languages and see if I could improve my overall skillset (and thus marketability). PHP is great for getting you started in creating a dynamic website. It's more friendly than some of the other environments like JSP, mod_perl (I'll explain), and more flexible (from my experience) than visual basic script + ASP. Also, it's easy to hook in and has excellent support from online hosting to the documentation on its website. Seems that the community for PHP is thriving as much of the newer more "hip" type of web utilities are being written in PHP. As long as you have some kind of database backend (namely Mysql or Postgres), it seems many of these utilities become a snap to just upload to your website and start using. Probably my biggest gripe is the sheer number of functions being used. I prefer Perl's method of using CPAN where you can customize the libraries by downloading what you need (which might be an incentive for me to move towards Gentoo as an OS). While it's convenient just calling a function name, some of the function names in PHP are just ridiculously long. And that makes them harder to remember. Also, I think it would be better if PHP starts implementing more object oriented methods in handling the more technical/sophisticated problems. I kinda hate having to do a ton of setup or even building my own objects to do the work that a good object oriented system could do. My other gripe is smaller but might impact people who run their own test servers at home then move them to an online hosting site. That is, versioning. Sometimes functions in PHP might be compatible between the minutest version change. Say 4.3.0 to 4.3.4. That might not be much but it can be annoying since it's likely that the people using the webserver are not the sysadmins themselves. That's kinda why I insist on the library/OOP idea so that people can install locally the libraries/objects they need. Perhaps they could have a repository for the general mass. Then if a user needs a particular function, they could place it in their own directory and modify the search path. Something I'd like to see as well is the equivalent of a DBI, connection manager. I think most people build their own, but considering how useful such a function/utility is, you'd think by now the guys would incorporate some kind of interface layer. Since this module would be quite useful, I'd suggest incorporating it at the binary level where it loads drivers for each database type, making most SQL calls and connections transparent. I'm a huge fan of placeholders and would think that having such a module in place would remove the need for us to hand parse and correct the parameters in non-integer formats. The upcoming version 5 one I've seen looks a bit like Java with the interfaces and more object oriented style. I find this to be silly since you're basically avoiding the point of using PHP: fast, easy prototyping for the web. If you're going to use Java, use Java! The one thing I'm not happy about is the exception handling mechanism in the new version. Anyway, bottom line: not a bad language to get you started on the web. Needs to be more scalable and modular.

Form validation

Thu, 19 Feb 2004 23:53:40 -0700

Been thinking of the most efficient way to perform a form validation using the MVC method for web development. Don't think there's a clear way since you have to validate based on the action. So you'd map an action to a group of items that require validation. Configuration seems like a bitch. XML schema was suggested by one person but sounds too slow because of loading the parser. Who knows? Maybe I should try it out. Want to make a module that handles the validation before hitting the database. Think it's better just to have the business method handle the validation since it knows what needs to be updated/added/deleted. Don't want to end up writing some crappy hack. Biggest problem is making things too complicated. Bunch of if statements seems like the most logical way.

Postgresql & Stored Procedures

Sat, 21 Feb 2004 12:28:45 -0700

I'm looking into how to use stored procedures in postgres, but at the moment I'm unconvinced that they're useful at this stage of implementation. I've done some stored procedure work in Oracle and a lot in Sybase. Sybase has multilevel transactions which is nice because of how you can rollback a transaction depending on the depth. However, postgres does not have this implemented yet. Also, postgres, at least from my understanding, does not have a way to capture exceptions. One nice feature in Sybase is error capturing. For instance, let's say you have this block: begin transaction insert into crap(field1, field2, field3) values (@field1, @field2, @field3) if (@@errors <> 0) begin rollback transaction @msg = 'Cannot add' return 0 end commit transaction With this, you can detect for the error after a critical operation like an insert or update. I've not been able to locate any documentation demonstrating support for this in postgres. This is REALLY annoying because it leaves me unconvinced to utilize postgres' Stored Procedure mechanisms. At this point, it seems that with postgres, it's better to let the application handle the transaction logic. The other thing is that I've found little support in terms of getting some programming languages to be able to use Stored Procedures with postgres. Perl's DBD::Pg mentions that it doesn't handle stored procedures while I've found no documentation of how to call a stored procedure from PHP. For Java, I'd prefer just to avoid stored procedures and do all my data logic with a middleware platform like EJBs to handle my transactions. Anyway really disappointing since supposedly stored procedures helps improve performance. Regardless, I hope what this means is that in a (near) future release of postgres (8.0?) that they begin supporting multilevel transactions and better error handling.

Might and Magic 8 & 3DO

Sun, 29 Feb 2004 01:27:13 -0700

Been off playing MM8. Very addictive but also very buggy. Shame that 3DO went bankrupt because they had made some really nifty games over the years. Been a fan of the MM series since I first purchased it on my Commodore 128. Geeze. We've all been around that long ;)

JRun

Tue, 09 Mar 2004 00:31:13 -0700

EJBs and JRun don't seem to mix very well. Got some stuff at work which had plenty of problems deploying beans. The real issue is that Macromedia's own book doesn't have much information on deploying EJBs. It seems more of a web container than a true J2EE solution. Crying shame considering that I wanted to work with it a bit. For the price, I think you might as well get the JBoss solution. There's enough complaints, I mean, documentation on the web finally that makes the container actually useable.

Sun & J2EE/EJBs

Sun, 14 Mar 2004 14:03:20 -0700

There's an interesting article I pulled up today regarding someone's rants on EJBs. It's an interesting read because it demonstrates many common frustrations in a well thought out critique of what I think is a massive hack. Some of the most salient points in the article I've found are:

Lack of portability between vendor specific software; you basically can't because each one has their own way of implementing the EJB spec and thus introduces a dozen or so problems
Lack of standards in deploying an EJB. I couldn't agree more with this one. I've ranted previously about JRun's deployment method. There's an IDE mechanism through a web based GUI but everytime I handled it, the internal mechanism simply failed. When I consulted all the manuals on the internet, Macromedia's online reference, and a book I purchased from the company on the subject, NONE had any solutions for deploying EJBs. This was tremendously disappointing.

I use JBoss at home because of cost, but the deployment mechanism is much simpler (just copy your jars, wars, and ears to the deploy directory and let JBoss do the rest). Still it's incredibly frustrating that this portion is considered non-standard. I imagine that in a future revision, this portion will become standardized, just like when the deployment descriptor did in the 1.1+ specs.
Auto generated keys. It's well known that many databases have various mechanisms for handling keys; but EJBs don't specify any good method. I'm surprised that this wasn't accepted.
Managing status lookups/enumerated type situations. From what I've seen there's no way to utilize the CMP 2.0 spec to handle this situation. In other words, when you have a foreign key in the primary table that references some static data in another table, how do you deal with it? The Wrox publisher's Professional EJB book mentions to use Bean Managed Persistence. This is absolute shit! What if you have a major object that requires to be referenced by other CMP objects? This is a total cop out! The other only way I see in handling this is to write a local ejb and use a session table to manage the relationships. But you lose the benefit of multiplicity from CMP 2.0. Did Sun even bother to think this portion out?

These are my biggest beefs with EJBs. However, my interpretation fo the situation is that you got these loud mouthed Sun marketing folks who know jackshit about programming throwing shit out their ass while the developers in-house probably do recognize these problems and took a phased approach (hence things like 2.1 specs). The problem is that the situation is just a huge undertaking so they probably got something out the door in terms of a spec and cop'd out, letting vendors to handle their bullshit. However, if you look at the situation, the whole thing is an under-thought out hack. Seriously. There's so much missing. Then things get lumped together even more because people are too afraid of backwards compatibility issues. I look at the idea of EJBs as Sun's way of blasting Oracle; it's a political strategy to move the emphasis away from the databases to the Middleware ground. I mean, the idea is to essentially let the container handle everything. The database is just a blackbox to store data. But the Middleware becomes this overly expensive piece of clump that essentially is the interface for your data. If you look at how they did this with the idea of "load balancing" and "clustering" how different is this from a highly available database? Then look at the idea of EJB QL. This is SQL! The idea of EJBs was to supposedly minimize the amount of SQL code the developers had to work with so they'd have more time to focus on the business aspects of application development. Now, they're essentially doubling their work with XML files, building queries, etc. If you're really worried about impacting performance on the database, why not just offload the middleware portion to a proxy like mysql and have it act as a middleware agent to your database warehouse? In complex applications, especially web applications which get hit by heavy net traffic, you still need something to pull that information out of the database. EJBs don't seem to be that solution. They seem to suggest using a combination of session beans and bean managed persistent entity beans for this. But essentially you're introducing far more complexity in your application and dependencies by having these layers, as opposed to writing a simple data class that handles customize database calls. I think the basic thing is that Sun tried to make a stab at this whole web/database market with their J2EE spec. The biggest problem is that it simply fails to deliver, imo. It's complex, fat, expensive and slow. Just like how Java was what Sun wanted the web to be, J2EE is Sun's attempt at becoming the big player in this ground. But I think it'll fail in the long term as companies seek to reduce cost, increase efficiency and begin noticing things like LAMP (linux + apache + mysql + php) which isn't as elegant, but seems to be more result oriented. That, in my experiences, is what companies really desire; not some bullshit promise that Sun is failing to deliver. It's a shame that so many companies bought into this scheme. I suppose it does have a place and the basic idea is good; it really needs to be fleshed out. The ultimate goals should be cost cutting, low learning curve, fast (development time, deployment time, and actual performance) and most of all simplicity.

Gomi Computers Are Great

Mon, 19 Apr 2004 22:25:20 -0600

One of my hobbies is to take a cheap, hand built system and installing linux on it. In the future, I hope to build a cheaply made data center from my old systems as a proof-of-concept thing. The thing is that you can get pretty good performance from a linux system which isn't too great. I'm thinking of building some web servers on low spec machines for an intranet while having mysql run on them. LAMP.

Mozilla

Sat, 01 May 2004 09:20:41 -0600

I love this browser. It's what IE should be. But one thing I'm a little miffed about is performance. At least with regards to my Linux desktop. I have a system with a 2.0 ghz celeron processer, 1 gb of memory, and 80 gb of hard drive space running Fedora Core 1. Mozilla sometimes feels really slow, especially when I use multiple tabs. Is it flash that consumes a great deal of resources? I noticed that this thing can be a memory hog. For some improvements on the browser, I think the Mozilla team should really aim at leaning out the memory usage and focus on the performance aspects as well as stability when used in conjunction with Flash. Since a great deal of the internet now is supporting flash animations, it really can drive the resouces on a system up. I seem to notice that a great deal of my crashes occur at the expense of some flash site. Hopefully, they can work on these aspects. Some people may suggest the Phoenix browser. Trust me, I use that too but I'm a bigger fan of the integrated solution. I find the stripped down browser to be too lean. Haven't tried other ones like Opera or Konquer though.

Heroes of Might and Magic 4

Mon, 10 May 2004 01:36:37 -0600

Okay this should be a dead stick as 3DO is gone. However, I recently picked this one up at Akihabara and have been chugging along here. It's basically another rehash of the three other previous games with better graphics, some tweaks in the game play, but nothing notably special. Actually, I've found the game play to be torturous in terms of time consuming. For whatever reason, these guys felt that the computer needed to outwardly display each moves it mades...including the monsters! This sucks because it's a waste of time. What's worse (for me at least) is that my version is in Japanese! Okay, I really wanted to play it but I couldn't find any retail English version while I was in America (I had two weeks and had many other important things on my mind). If I had a major complaint about these two series, it's that they rewrite the games over and over again, but end up returning to the same formula. I guess they want to maintain what they've done over the years to keep their following going. However, the worlds themselves never really evolved outside of the graphics, imo. Then again when I saw the "newer" version of the old Wasteland game (I think it was Biohazard or something), they practically did the same thing. Maybe this is why 3DO went out of business....

Linux Direction

Sat, 15 May 2004 10:36:54 -0600

I enjoy the fact that Linux is becoming a major competitor to M$. However, I think a huge problem is getting more major software companies to adapt/port over their applications to the system. A few major ones I think would benefit the Linux system are Adobe Photoshop/Illustrator, Macromedia's Dreamweaver, and Intuit's TurboTax and Quicken. I find that those along with games are the few things that make me still use a Windows computer (although in my case, I really only use Photoshop and TurboTax). However, I think many PC users might consider switching over to Linux if a lower cost Photoshop was provided. I've read various rants how those particular software items are the ones preventing permanent move over for Linux users. The thing though is that for Linux to make a breakthrough in the consumer market, it needs to grow in what consumers perceive as necessary applications like those mentioned above. The great thing is that other major applications like OpenOffice, Mozilla, etc. are already provided (for free!) for the Linux operating system. Next will come brand awareness, stability and increased compatibility for adoption. I think RedHat, Novell, IBM, and some other major players in this market should make some major pushes for this desktop market. There's some good efforts in place for the open source market for similar products like Gimp, but hardcore Photoshop users might not feel the sheer power in the application. For Linux's success on the desktop market, imo, it's need the equivalent of a Windows '95 milestone; in other words, the point in time when most consumers are willing to adapt it because they see it as a real need. But it's really coming close....

Fedora Core 2

Thu, 27 May 2004 23:36:08 -0600

Had a brief chance to play with it. Much faster! There's a network problem with IPv6 but I also have solution for that in the Tech Solve section. Thinking about upgrading my workstation at home but not sure of the technical details. KDE looks very good. Waiting to try the new Gnome interface as well. Haven't used it enough to judge what other improvements have been made. Can't wait to upgrade my other systems!

next html version

Sun, 30 May 2004 22:59:36 -0600

my wishes for whatever the next version of html there is: 1) curved and angled table ends. this seems to be an extremely common practice these days where most web designers the photoshop to create curved or angled edges for tables to give it a more "professional" look. I think it might just be easier if HTML can render those things by themselves. i think it would save valuable time just by specifying the end shape of a table attribute and dumping the look into a css stylesheet rather than implementing an image. 2) larger variety of fonts with more depth. things like drop shadow, color specification (like outer color and inner color to give a glow effect). currently, what exist is very limited. 3) more widgets. not just buttons but say tabs that are customizable using stylesheets, or rounded pills. have these be substitutable for the banal hyperlink and images. 4) in general, anything that's being in photoshop and ported over through excessive tweaking in html, should be rendered through html. that would, imo, alleviate more rendering problems since it's quite difficult to get browser for different operating systems for different resolutions and video cards to adjust for the current problems of meshing pictures and html. let pictures be pictures and give the browser the bigger task of rendering things.

Fedora Core 2 Bitch

Wed, 16 Jun 2004 01:01:38 -0600

Okay, I've installed Fedora Core 2 on another machine that is for my main workstation at home. Some stuff works much faster. Some of the load times are better. The GUI looks slightly more polished. However, I'm somewhat insulted they called this a "stable" release. 1) There has been a noted issue that sound cards (Sound Blaster!!!) doesn't detect from the outset. That's definitely true for me. It's probably because I have an onboard sound card and need to disable it. However, the onboard one while detecting at the beginning, doesn't register now. I just hear crackling. Really pisses me off that something so basic could be overlooked. 2) Networking is fucked up. In my Tech Solve, I point out a setting that needs to be corrected in modprobe as a result of the move to IPv6. Shouldn't this issue be handled on first install? You can't even use Mozilla properly with a simple install. Very disappointing, especially for novice users. How can the Fedora Core community convert more people if their stuff is broken? 3) Dual boot problems. I don't have this problem because I don't dual boot. But shouldn't an issue like this be moot by the time they declare it production ready? These have been the big killers for me thus far. Also, I've found Mozilla 1.7rc3 to be slower. Not sure if it's because of the Macromedia Flash Plugin. But response times for non-flash pages has been pitiful. There's a dramatic pause before some pages load for about 5 seconds or so. Not sure if this is an issue with the OS or the browser. The plus sides has been my graphics card detection along with dual monitor. This is awesome. I've got the dual monitor working whereas Fedora 1 couldn't handle it. So I can do twice the work with my second monitor. Very sweet. Haven't tried printing yet. I'm interested in seeing if they've updated their drivers since my Epson printer last time had issues since the driver wasn't supported. Guess I gotta mess with it more. But most of these issues should be handled by default after the install.

Forget Fedora Core 2

Sun, 20 Jun 2004 12:54:43 -0600

Okay, I've toyed enough with this piece of garbage long enough. It seems like they didn't bother doing enough testing before getting this crap out the door. Although it did fix my problem with the dual monitors, there's just too many other frustrating problems to really make this distribution useable. The sound problem is nasty and the networking issue is plain idiotic. For a hobbyist or a useable destktop workhorse, this isn't ready for production. I think they should just cancel this as a "stable" release and wait until the next generation of the 2.6 Kernel improves along with fixing all these problems. No doubt that the staff at Fedora know about these problems and will amend it. Personally, I hope they will avoid trying to put too many new features in and just focus on stabilizing the main distribution. In the meantime, I'm going to rollback to Fedora Core 1 since that was working just fine.

Rolled Back to 1

Sun, 20 Jun 2004 17:58:23 -0600

Got back to Fedora Core 1. Sigh. Very disappointing since the new features for Core 2 looked exciting and promising. But I can't live without sound and there's too many undocumented hacks for amending it. If I were the core team, I'd make an intermediate release to try and fix these problems asap. One thing I noticed about Core 1 (I had to download it again), is how out-of-date the binaries are. I think they should be putting out maintenance releases with the updated binaries rather than having people do fresh installs and performing massive upgrades. It also disappoints me that some of the newer desktop themes aren't prepackaged for older distributions; yeah, if you really want it, you should compile it yourself kind of thing, but there should be some kind of limits on doing versioning and compatiblity. Say that for two years, they'll upgrade releases to be compatible with older systems (since as demonstrated with Core 2, you don't always get the most reliable system with the latest release). That way both parties can be happy. I suspect that they'll have to work closer with the kernel developers. For instance, the solution supposedly for fixing the Core 2 audio problem was to shut off kdearts (or something like that)! That to me is ridiculous as a stupid hack. These guys really need to get their act together soon for this shit to start working and be more acceptable for the public. For myself, I love the idea of a community oriented linux system. But the latest release puts a very bad taste in my mouth. I'm someone that doesn't have enough time to screw around with documents and config files. I have it bad enough just getting crap like jboss interoperating with Apache and php, etc. Why fuck around with my desktop? Fortunately, my desktop is pretty vanilla with the exception of a few utilities and the gratuitous female background decor. Still, when I gotta get it working, I have no time for major problems. Fortunately for Redhat, this is going directly under their name; it's mostly research for them. However, if they are going to be a key player in this, they have to realize that for the enterprise, pushing crap out just for the sake of rushing a product out is just horseshit. I'm patient and don't mind waiting for a stable release. But it's inexcusable for me wasting a few days getting crap like this working and then rolling back.

Postgresql 7.5

Fri, 02 Jul 2004 22:42:26 -0600

There was some interesting news regarding the upcoming postgresql 7.5 release. first, fujitsu supposedly has put in a good amount of funding to help them handle multilevel transaction (something avid readers of my blog probably notes is a huge bitch with me and the current version of postgres). Not sure what else is happening but heard that some enterprise level features are the works. looking forward to it since postgresql is basically a good open source replacement for industrial strength DBs like Oracle.

Sun to buy out Novell?

Tue, 03 Aug 2004 01:28:51 -0600

There's some talk on the street about a possible acquisition by Sun to take over Novell. Novell's real position has been it's recent own acquisition of Germany's SuSe Linux which has become one of the highest supported Linux platforms next to Redhat on that market. However, Sun has done some partnership with Redhat to sell it's own version of a Linux based server, which utilized AMD technology rather than its own sparc processors. Many of the other major software and hardware companies have recovered somewhat from the fallout of the dot com/technology fall out of the late 90's. Yet Sun still struggles to regain it's footing. Part of the problems exist in the fact that Sun has not shown any keen devotion to one particular strategy to propel itself forward. Some days they're pushing Linux. Some days they condemn it in going with SCO. Whichever way, probably the biggest battle is in its decision on Linux. There has been tons of criticisms of the Solaris OS. Add to that the immense cost of Sun's hardware, where you virtually lock yourself into a vendor and you end up limiting what you can do. This is where Linux's penetration into the market has allowed more hardware oriented supporters/makers like IBM, Dell, HP, etc. to focus on diversifying their products and allowing the market to be truly competitive. However, Sun has remained noncommital about their role in all of this. That's mostly an internal problem with Sun and their politics, probably akin to how a company like Sony faces internal legal battles between their electronics division and their entertainment division. In order for one sector to make money, the other side must sacrifice their products. Not a pretty role to be in. But what if Sun makes an acquisition? What is it really worth? Novell has pretty much died down as this hoopla for networking since the days of Windows NT, etc. However, the move to buy out Suse was a great one since Suse controls a good portion of their market. Redhat has always been a bitter friend for Sun since essentially it nips away at Sun's Solaris OS. If Sun purchases Suse then, would they keep on supporting Solaris or would Suse simply die out? I don't see them buying out Redhat in all of this. The personality doesn't seem to fit the bill. I think Redhat ought to expand it's software business from just OS products to certifying more large scale open source type of products for medium to large scale companies. That way they can increase their revenue without relying just on their OS and support for their main revenue. But Novell is a tough one. Sun does need a major kicking from some place so this might be the area. But I'd hate to see Suse die off if it becomes another internal war in Sun.

Lycos Sold

Tue, 03 Aug 2004 01:39:50 -0600

Apparently, Lycos was sold by Terra Networks to South Korean Web site operator Daum Communications for $95 million. Probably one of the few global portal sites left that was competing against Yahoo on it's own merits, Lycos seems to now try to differentiate its services from Yahoo, although the basics remain similar. The thing is that Lycos is far scaled down version. But like the strip in Las Vegas, Lycos is able to maintain some distinction with a few of its branded services like Wired News and Matchmaker.com. Still Yahoo is just so vast people still think *it is* the internet (although google instinguishably might get that title reign in the next few years). So the question remains how will this acquisition change Lycos if any? I think Lycos should continue on creating branded services that they acquire or create rather than compete on the portal/search engine layer.

EJB 3.0

Sun, 08 Aug 2004 15:12:57 -0600

Just saw some news on EJB 3.0. Hard to say if I'll like it or not because some of the things they are using, which I am not familiar. But there seems to be some positives coming up such as the elimination of deployment descriptors (just a mess which is error prone!), elimination of home interfaces (great decision as it was just a useless class that added extra verboseness to an already verbose system), and the removal of session bean interfaces. Instead for the session bean interfaces, they will be moving to an "annotated" style. Also, JNDI will be removed in favor dependency interjection (why create such a mess of a name?) In general, it seems like the general goal is to make EJBs more lightweight or more appropriately just more friendly to program. For myself, I use local entity beans (which basically all represent a row in a database) which are accessible through remote session beans. Sounds silly considering they suggest that entity beans should not necessarily represent a row in a database, especially with the advent of relationships, but i still don't understand the whole relationship bit. I think that's horseshit because the relationship system in EJB 2.0+ is too limited. Complex relationships then move to BMP rather than CMP situations, which just is fucked because it defeats the whole purpose of relying on the container to manage things smartly. I like the idea of declarative programming instead. I don't mind a system that could utilize an XML type format (not that horrible deployment descriptor though!) to describe the relationships intelligently. The container should then generate an on-the-fly type of object based on your call (and it would be of four types: select, insert, update and delete). Hopefully, part of this would eliminate the need for value objects (which supposedly local entities were supposed to do, but I'm not buying that) as the object would be dynamically generated as a hashmap type of object and the data values would be extracted as key/value pairs. Basically, we just want something to map our database with customized settings for our activities. EJB provides half that solution with an interface and layer into the database that standardizes how that interaction should occur. However, EJBs are simply too burdensome. Of course, I haven't used many code generation tools like Xdoclet (no time), but even then it takes too long to design. Not surprisingly it seems like people are returning to perl or even php for web design stuff since prototyping and rapid development is better for those platforms. But what can you expect?

Epson Printers

Sun, 15 Aug 2004 17:10:22 -0600

I utterly trashed my printer today. The stupid thing jammed (again!) just on an insertion of printer paper. It wouldn't recover so I got so pissed, I literally tossed it down several times. I'll never buy that trash again. I've had so much better luck with HP.

Class::DBI

Fri, 17 Sep 2004 02:36:21 -0600

Started working with Class::DBI. Was a major pain to get working. Don't know why the author decided not to make this more object oriented at least in the typical way (e.g. new method, inheritable). I think you can but I must be missing something. Trying getting this other Class::DBI::Loader working. Too many problems. Seems really broken. Then again it didn't support what I wanted which was an easy way to submit a dbh. Instead this family requires a DSN type of string. Very annoying because I wanted to re-use some established dbh objects I prepopulate. Ugh.

Jonathan Schwartz

Mon, 24 Jan 2005 12:17:44 -0700

Man, the more I read about this slob, the more he reminds me of some punk that I'd love to get on the street or a UFC ring. While I'm certain he's not the sole cause of Sun's decline in recent years, he's definitely one of the key malignancies with that company. I've yet to hear one useful word from him or one vision which can take Sun back to prominence. Frankly, given history I see it as an impossible feat. His recent open letter to IBM reeks of desperation for more companies to embrace Sun. IBM should simply buy them out already and layoff everyone except the engineers; Oracle did that to PeopleSoft, so IBM should make the next move here. Except I think IBM is smarter than that and realize that Sun is dying a prolonged, remorseless demise. Apple managed to reinvent itself with iTunes, but Sun can't compete in the server market these days with it's late Linux strategy. Either way, I don't think this guy has a good vision of the future of technology for his company. It's too late for Sun to regress and focus on the server market (which really their niche). I think Sun should simply layoff its sales engineering, marketing and upper management. Just because you're young, have glasses and believe you kick ass doesn't mean that you don't deserve your ass kicked up and down the street.

Developers...developers...developers

Thu, 27 Jan 2005 04:54:53 -0700

With all the buzz about C# vs Java vs Python vs Perl vs PHP, et al, I thought about the relationships between the above in terms of how it applies to companies and employment. First, I want to say that I don't believe any single language is better than the other as defined within a certain scope. For instance, Java vs Perl. The issue at stake wouldn't be the nature of the language, but whether the language is suited for the job. People being religious about those issues are being immature in evaluating the purpose behind a language; part of that involves just how competent a person is within that language. It's been interesting seeing how some companies have made a huge turn around depending on the system they're using. For instance, Friendster going towards PHP while Ebay moving towards J2EE. My belief in those cases is that the language/platform didn't make the difference; the difference was in the people implementing those systems. After talking to various people about how their large scale systems managed to work with say mod_perl and no middleware layer, I became convinced that the J2EE/EJB/middleware system being this wholistic solution was just a FUD by Sun/Weblogic to convince people to buy into their systems. Why is this? Well, the real issue at stake is the core competency of the team involved. A good system architect matched with a team who share similar philosophies can outmatch these so-called packaged solutions because they know what they're doing. They have a strong grasp of development and the technology they're using, so implementing business requirements becomes second nature. Unfortunately in the tech world, good technologies tend to be undermined by FUDs created by marketing/sales and recruiters as well as the biases of the people in charge. The whole dropping of mod_perl was quite sad; part of that seemed due to J2EE getting a lot of hype with places like BEA talking to the business rather than the engineers. Of course you had other reasons like PHP gaining steam and python coming as a response to the inherent problems in perl. But I've seen places where a few good engineers could make any of those technologies work (as well as where those technologies could seriously fail).

More PHP

Wed, 02 Feb 2005 19:10:28 -0700

Been doing a ton of coding with PHP as you may have noticed with all the updates on the site. Got my first opportunity to use PEAR. the command line edition seems simple and comparable to apt-get. however, apt-get is far more intelligent since it will install the identified failed dependencies. Started working with some of PHP's XML libraries. Not too bad. The new OOP interface is far better than the functional one that's been deprecated. Funny thing is that there's no backward compatibility. I don't mind PHP as much these days but the changes between versions are too large, making the API kinda unstable.

Sun: All Your Solutions Belong To Us!

Sun, 06 Feb 2005 23:56:43 -0700

Well, that's according to their little slideshow from McNealy up on cnet.com. Now they want to get into the database market. Their view is literally the Sun stack. It seems funny that they want to invest in this market where there's so many players far ahead in the game. On top of that, they supposedly are looking into the open source database market as another strategy. But I think that's the core problem right there: they're trying to get into the game and brand something that's really not a specialty of theirs. Here Sun is overextending themselves and trying to be an all-in-one solutions company. When you look around them, it doesn't seem to be a real success. We see IBM sell off some of their hardware components like their notebook and hard drive businesses recently. Dell has never gotten directly into the software component and remains committed more to reselling outsourced components that are QA'd and certified with their branding. HP has a bigger inventory and makes larger commitments with even having their own OS and software division, but they still by far and large are a hardware manufacturer. In going into the DB business, we see Sun spreading themselves too thin especially against heavy competition with Sybase, Oracle, Microsoft and IBM as their chief bigname rivals while Mysql and Postgresql are growing at exponential rates in the open source market. To me, this is a critical, horrible mistake that will cost Sun billions of wasted dollars since most customers won't want to commit and migrate to yet another platform. Part of the issue I see is the whole branding concept. Their so-called invasion into the linux world has been a dismal failure as they simply re-branded Redhat linux to be certified with their hardware. Again like I said about Sun, it's another one of those "let's toss something against the wall and see if it works." By examining the failures of a company like WCW in that method, they should understand that it's a poor method of doing business. Sun's greatest three assets are their hardware (mainly their sparc systems), Java (mostly the concept as opposed to the products they tote outside of the JDKs and JVMs), and Sun OS. Branching otu limits the focus and spending they can achieve on their core competency. They could've used this money for increasing the performance of their sparc processors while lowering the cost. While businesses may not want to disentangle themselves from legacy database systems, they certainly would be enticed by a lower TCO for the hardware portions that already exist in their data centers. If the long term strategy is to become a wholistic service provider, Sun might be really out of their league at this stage of the game. Taking one look at the Sun stack, you should automatically notice that there simply are too many layers. The middleware layer itself adds three layers to ones TCO by demanding people to pay for the software, hardware and maintenance/development costs for it. Also, the DB layer most likely will involve a good charge since it'll be Sun branded. A LAMP product, on the other hand, decimates the TCO against Sun; you have three layers in LAMP for cost: the hardware layer and the developer/maintenance cost. With Java developers, the cost is at a premium compared to a good LAMP developer; also, it has been well documented that the turnaround rate for LAMP style projects kill Java out of the water because of the rapid prototyping model that exist around perl/python/php + postgres/mysql. You're essentially saving yourself 1/3rd the cost of a Sun based solution just on the number of components you'd elimate. Sun might argue against scalability in Java vs the LAMP model, but that argument is quite fallacious for all but the largest enterprises. And even then there's enough proof that LAMP can scale up against Java just as Java can scale down in an inappropriate situation. Again, this is another desperation Sun move to buy into their wholistic solution schema. What Sun constantly fails to realize is not putting out these pretty slides for customers to buy into but that they should simply focus on making better products at more reasonable cost. Part of what Sun fails to realize (or maybe they do but refuse to admit it) is that complexity is really detrimental to their business model. They seek to add more costly components so they can increase business. However, the thing is that businesses (and people) like simply, fast solutions that are cost effective. If you look at the direction of computing, it's interesting to see the rise of scripting languages as being the prominent tools for many developers. You can't deny the power of tools like perl, php, html and python in the evolution of the web. Why did they grow to be popular? Well, it's the same reason why VB was popular with businesses: sheer simplicity. Scientific/academic computing is best left to its namesake. However, the dead truth is that people are lazy; we needs tools that decrease the complexity of our lives, not increase it to sell a fast buck. I think that Java, although having a nice support community, is something that confounds itself because of the sheer complexity involved. Sun needs to cut the crap with it and come forth in producing less restrictive software standards.

FC3

Tue, 08 Feb 2005 23:58:01 -0700

Got to try this out finally. Nothing spectacular that'll make me upgrade my home distribution, but it's nice having the latest packages. I think the next big change is coming with FC4 so I'll wait for that. I'm guessing FC4 will introduce Gnome 2.10, KDE 3.4 and the most important development, Open Office 2.0. By then I hope they resolve some of the sound problems with the integrated alsalib in the 2.6 kernal. But FC4 = RH9.2. My guess is that FC6 will be the next major switch.

on top!

Sun, 13 Feb 2005 11:42:29 -0700

yes! now when you do a google search on my name, this page returns as the first result. I rule!

C++ Programming

Wed, 16 Feb 2005 09:28:27 -0700

Been doing some C++ programming at work just to continue practicing. Not good by any length of imagination, but I wanted to try it out a bit. Can't think of any useful projects to write in it though. I'm not even sure if it's something I really want to tackle at this stage in my career. I think most things are moving towards Java and scripting languages. Want to try out the Perl QT module. Not sure if it's more useful than the Java swing part.

NY Times buys out About.com

Thu, 17 Feb 2005 23:52:46 -0700

Purchased this from Primemedia for $410 million. About.com is one of the highest ranking sites out there in terms of sheer traffic. Basically, it has various "experts" who write up on various topics. Sorta like a more in-depth portal rather than a plain directory. Most of their money is derived from online advertisement. However, user retention is low compared to other high volume web portals. Despite having such a high volume of users, I find the purchase to be a waste of money for NY Times. The big thing for them is the merging of hard content with online content. Still I've never really been a big fan of their content. There's too much to sipher through and I'd prefer just googling around to find the real result I want. Probably because they've been around for so long, they've managed to acquire a large population base. However, there's plenty of better websites that produce more focused content.

Firefox

Fri, 18 Feb 2005 23:25:01 -0700

This is the most kickass browser in existence. sure there's not many, but it being open source with an API allowing other developers to pitch in, you get some killer plugins. I think mozilla is going to be phased out slowly in favor of the split suite combo. Heck, the development on Mozilla has been ass-slow. they used to burn out new versions every month, but now with firefox in production with 25 million downloads, I think the developers are going to be more focused on that and Thunderbird. I use both btw and I'm extremely happy with them. if you haven't downloaded firefox, you better grab it now! the longer you use any other browser, the more you're bound to infect your system with some form of spyware.

gmail account

Sat, 19 Feb 2005 10:45:30 -0700

just got a new gmail account. probably will use it for spam and testing purposes.

Follow the Leader

Sun, 20 Feb 2005 22:04:16 -0700

Read how CA is planning to release some of their software patents after IBM and Sun made their moves. Of these guys, IBM is the only one that has a real strategy. Sun and CA are following IBM's move towards Open Source but their not following IBM's spirit. I believe IBM's moves towards supporting the Open Source community is a major assault against Microsoft's position as well as having the fringe benefit of increased participation, innovation and community favoritism. IBM is doing the smart thing by leading by example. I seriously doubt you'll see Microsoft following suit as their strategy is just the opposite. Still it'll be interesting to see what other companies might follow. I can't see Oracle nor Sybase moving in this direction. HP might do it though since they have a major Open Source initiative. I also doubt the game industry moving in this mode; companies like EA really have no rivals in Open Source since games are almost exclusively developed for the Windows and Mac platforms. But it seems that we might be seeing more business oriented software go towards this area. Just a few weeks ago, JP Morgan made their announcement to contribute an Open Source messaging system. Quite interesting considering this is a major bank.

Eclipse

Wed, 23 Feb 2005 08:39:49 -0700

Well, that's what Sun is doing to their 3600 developers in an announced massive layoff to return to profitability. Not coincidentally our friend Schwartz got promoted the same day the initial 3300 were announced. Hmmmmm....do you think that his number of shares increased with that promotion? I still think those numbers should've been adjusted towards their sales and marketing department. The software people are the good guys in all of this. I would've cut a bunch of their sales and marketing directors and middle management for those sections rather than the software engineers since those are the ones actually creating the products.

M$ stepping up anti-piracy efforts

Sun, 27 Feb 2005 13:23:46 -0700

now supposedly PC owners have to call M$' support if they want to activate their PCs after a re-install starting this March. as part of anti-competitive measures, the government should impose a sanction that forces all retail outfits to carry a large sticker saying this is part of the bad deal you're getting everytime you pick up a Winbloze PC. along the opposite lines, Redhat, Suse/Novell, etc. should start a marketing campaign to promote Winbloze awareness. This is such an easy target for Linux to gain market share in the desktop market. Here's the commercial: some guy picks up spyware through MSIE that installs a ton of p0rn leading to his wife divorcing him. As he's doing his taxfilings at the end of the year, his PC's drive dies so he has to reinstall Winbloze. he waits and waits and waits....and then it's too late for tax filing so he's screwed three times over! Pan to his neighbor who's cruising on his Linux box with KDE and browsing the web safely via firefox. In the end, he nets a huge return on his taxes while his neighbor kicks his PC out the window. The tagline should be: "The Choice is Yours."

Why Apple Isn't The Good Guy After All

Sat, 05 Mar 2005 00:30:29 -0700

Apple like the RIAA are going after people with lawsuits in the crack down of their trade secrets being distributed to web publishers. Apparently, three sites just got hold of an upcoming music player Apple is working on and Apple like the Salem Witch Hunt has been interrogating employees to find the insider. Funny if you view the whole Pirates of Silicon Valley, you'd see a different beginning from the humble disjointed start of a couple of hippies trying to defeat IBM. Look at how the tied has turned! Nowi, it's Apple aligning themselves with the new face of evil: lawyers and IP while IBM is going forth RELEASING patents. Well the only real visionary in the company was the Other Steve and he left and is doing real social service. Apple is left as nothing more than a greedy company with their finger shoved in their ass. Never liked them, never will no matter what people say. Think upon that the next time you look at an ipod or ibook.

Got a new gateway

Sun, 06 Mar 2005 04:34:04 -0700

3.2 ghz P4, 1 gb ramm, 200 gb hd, dvd+/-. too bad the graphics card sucks. had some issues with flickering with my monitor (a Go Video TV/combo). good for a game machine though. Looking forward to playing a few games like Neverwinter Nights and Pool of Radiance.

Firefox Pissed Me Off

Sat, 19 Mar 2005 10:59:41 -0700

The fucking backspace (which we commonly use for DELETE!!!!) did a backpage after I was writing a blog. You fucking MORONS! Don't you know anything about UI??????? Anyway, get the keybinding extension: http://dorando.emuverse.com/projects/mozilla/keyconfig.xpi and disable that stupid backspace key. (It should be disabled by default you idiots!!!!)

Absolutely Needed Firefox Plugin

Sun, 08 May 2005 22:49:29 -0600

Make something that recovers form information. Or at least save it into cache something in case someone's browser crashes. I get too many browser crashes where I'm typing in a form and lose all my information. This is just as bad as the MS Blue Screen o' Death syndrome. Not cool!

php and scripting languages growing in acceptance

Sun, 15 May 2005 22:24:23 -0600

cnet has this article: http://news.com.com/Grassroots+computing+languages+hit+the+big+time/2100-1007_3-5705448.html?tag=nefd.lede it's interesting to see the growth of PHP and other scripting languages. Perl is my personal shotgun but my alternative has become PHP (as you noticed by this website). but the thing about scripting languages most frequently boasted is the ease of programming. That leads to a high turnaround rate in projects since you don't have to handle some of the more pressing issues like garbage collection or the verbosity of a language like Java. Focusing on PHP for a moment, it's interesting to note the number of jobs that has popped up within the past few years. Just as an experiment, I did a quick search on Dice.com using as my criteria the number of jobs in LA within the past 30 days with just the keyword php. I pulled up 59 results. A few years ago, it was pretty tough finding any work for PHP whereas you'd see it more on the web hosting end. Now, the language has shot it to be taken more seriously by some industry leaders like yahoo, friendster, idealab, and now Oracle. Of course, because of the ease of use, the rate of a PHP developer still is lower than that for a mod_perl developer. But it's good to note that there are other opportunities in case the perl well dries up (not likely).

p2p applications

Sat, 21 May 2005 23:17:31 -0600

got off the phone with a friend in discussing how you could create a cheap, fast file sharing application. why create it? it already exist with p2p software. the issue is that we needed an infrastructure to support sharing of files in a folder where only certain users can access it. VPN is a good method since you get that secure layer. However, you have to take into consideration licensing and experience. on the other hand, P2P is quite efficient for smaller projects that need a quick and dirty setup. naturally the RIAA and MPAA, in their selfish quest for a total monopoly over production (read Marx), have decidedly voted against P2P as a viable technology application. however, let's examine it from a business perspective. compare this to CVS, subversion, samba or similar technologies. those were key considerations for the project since the system we originally proposed didn't seem to scale with the bandwidth requirements. while those areas are efficient in their modes, the average user must be trained in utilizing those areas and/or the infrastructure must be adequately setup for it. with a limited time constraint, there's no way we could produce this in a week's time. Enter P2P. is it the best method? probably not, but it's a good intermediate bandaid with real world business applications. training users for this shouldn't be an issue either since it's like a search over a network of folders. limiting this over a WAN through the net to only a few select users so that people can easily share files on a project definitely demonstrate key benefits of why this technology shouldn't be shunned on the basis of piracy.

Started Using PEAR

Sun, 22 May 2005 02:40:32 -0600

I did try PEAR for a couple of things before like XML and RSS, but now I'm using it with DB. It's basically PHP's DBI module, except less feature filled. It does take some time to get used to as it attempts to accomplish the same means as DBI, except with a parameter being moved here or there. For these jobs I always prefer DBI over most db interfaces since perl has some strong features that cut down the lines of code. But this isn't that bad because it supports placeholders. I'm still getting used to how it implements references. Kinda confusing at first. Haven't figured out as many shortcuts with it as I have with DBI. Also, I'm going to try smarty out. This one is PHP's version of Template Toolkit or HTML::Template. Funny that PHP is using another templating system to do code separation in an MVC type of setting. Like Template Toolkit, it supports some page caching which is a feature that HTML::Template lacks. Gotta tackle it some more to check out what other features are available. Seems like you can do some localization by utilizing config files on a per template basis. Can't say that I get a hardon from this but I can't say that I'm absolutely hateful of it either. Maybe I'm too professional at this point. Or maybe I'm just a language whore :p

Smarty

Mon, 23 May 2005 01:17:10 -0600

Not a great name for a system, nor is the most original, but it's actually fairly easy to get started immediately. been hacking around with it. At first, I was reluctant because I'm a cynic when it comes to PHP having another templating language do what it should be doing: being a web programming language. However, PHP requires a ton of setup if you want to create your own tags. I mean, you can't really do something like: < if check_block > you are an elite dude < /if > without hacking the system. fortunately, someone did hack the system and made it work like a template toolkit. that's where Smarty comes in for PHP. some nice features about Smarty are the following: 1) page compiling and caching. with your templates, Smarty will compile the output and store it on disk. With caching enabled, you should get an increase in performance since you won't have to regenerate the page each time. I don't employ Smarty's caching because of the way I structure my system and instead utilize some other caching layer for my pages. (Which has it's own issues, but those should be solved in due time) 2) Config files. I don't like the idea of them calling these "configs" because they really aren't. They're essentially like the old properties files for Java in that you set a string to a key and can access them through the template. They have two main strengths which is centralizing messages and more importantly providing a potential mechanism for translations of messages. This is incredibly useful if you plan on having a multilingual website where you can devote people to translate the keys. 3) built-in tags. they provide some standard mechanisms to help clean up your template code to very minimal logic like if-else statements, looping and, of course, variable substitution. this helps in avoiding the terribly ugly {} characters and php start lines floating all over the place. 4) plugins - like template toolkit, you can write your own plugins. and it's not that hard. in creating your own plugins, you can expand the tags to fit your needs. Overall, if you're anal about splitting your html code from your php code and need some kind of quick and dirty mechanism, give this a try.

Javascript and DOM

Sat, 28 May 2005 05:46:37 -0600

If you haven't discovered the latest capabilities of Javascript when combined with DOM, you're really missing out! The introduction of DOM into browsers and javascript really allows you to gain so much more dynamic control over your page. One of my favorite things is to generate additional form elements without resorting to a page reload. Combining this with CSS proves to be a powerful combination. For instance, within a div tag, if you have text you want to swap out, you can load it all into the page, and hide the text that should appear only when the user say mouses over a link. Or perhaps that information can be pulled from XML via the server and then loaded dynamically into the page without causing the page to refresh (this is also known as the AJAX method) Something I found to be particularly hideos and cool is when you utilize dynamically generated hidden input elements. With dynamically generated elements, there's no way the user can see via the view source option the base code of those elements. So in theory, you can pass in hidden items that truly are hidden except with someone who's extremely scrutinizing. Overall, very cool stuff.

sound card support for fedora

Tue, 14 Jun 2005 14:36:07 -0600

http://www.alsa-project.org/alsa-doc/doc-php/template.php?company=Creative+Labs&card=Sound+Blaster+Audigy2+ZS+Platinum+Pro.&chip=emu10k2%2C+CA0151%2FP16V&module=emu10k1

bizarre video

Tue, 14 Jun 2005 15:02:18 -0600

http://www.big-boys.com/articles/numanuma.html

external usb drive

Wed, 15 Jun 2005 02:28:25 -0600

got one the other day at fry's. getting ready to setup fedora core 4. haven't updated my OS since core 2 had sound problems with my onboard card. core 3 i heard still had the same issues so i decided to wait it out. so far so good. since this drive is bigger than my main drive, i'm thinking that i might start doing nightly backups (or daily since i'm never around). also, i intend to do move all my high volume information onto this disk and save room for more critical applications. saw a 1 terrabyte raid system. looked killer. cost a lot though, but it's something i wouldn't mind having in the future especially for keeping all my main data on storage. backing that up would suck though. next i gotta figure out how to perform nightly, incremental backups. i know rsync is capable of this, but i don't know how rsync could work on a local system. it would be far faster than a raw copy, but i need to perform some research on that. going back to fedora, my disks were ALL corrupted. maybe i burned them at too high of a speed. spent the rest of today reloading my old core 3 system. damn that sucked. hopefully, after backing up this batch and re-burning core 4, everything will run smoothly. my main concern at the moment is getting the sound card on core 4 to work. if that happens, then i'm pretty set.

FC4

Fri, 17 Jun 2005 11:13:42 -0600

Just a hint to all those out there upgrading their OS or intending to upgrade their OS when a new distro hits: wait two weeks for them to correct any problems then upgrade.

hold off on core 4

Mon, 20 Jun 2005 12:22:28 -0600

is it me or has core 1 been the only decent distribution coming from the Fedora community? i use core 3 at work but that's because i don't require sound on my workstation. core 2 failed for me because of the transition to alsalib (which supposedly has not been fixed so that you had to dive into the alsa console and unmute the volume settings). my little sharp laptop choked on core 3's installation perhaps because of the LCD screen (alignment was really messed up). for core 4 though, i went through two sets of disks burns and both came out with checksum errors. heard this occurred starting with core 3. this is NOT good. i appreciate the open source effort for Fedora, but i'm finding it hard to justify moving away from core 1 at home and core 3 at work when it seems as though more problems keep cropping up. after reading through the mailing list, it seems there are issues even in the installation process. this looks quite shabby for such a promising distribution.

DoubleShit FUDing

Thu, 23 Jun 2005 13:18:39 -0600

DoubleClick's CEO is sending out a FUD on the end of the free internet because of ad blocking software like Firefox's ad blocker plugin. Uh, I doubt that I ever clicked on a DoubleShit ad in my life conscientiously because their stuff sucks. This guy is completely deluded and self centered if he believes his company's obnoxious "solutions" improve our browsing experience and don't account for a large portion of the spyware and viruses perpetuated out there by machinations of whoring themselves out. Good, old text ads like Google (click on top!) are what I like. Not these popups, flash crap, etc. that drain bandwidth, resources and are intrusive to one's privacy. Don't forget the ever sneaky things like double embedding javascript with rewrite tags to force popups to appear. Keep it simple and effective. Either way, I won't be crying if DoubleClick files for bankruptcy sometime (or their parent company). Google is enough to keep us happy :)

The internet

Mon, 11 Jul 2005 00:38:35 -0600

I just talked about progress and one thing that I think needs to progress is the technology behind the internet. A lot of these venture capital groups went crazy when the ebays, amazons, yahoos and now googles went IPO. And so did the job market. But in recent years there's been such a huge emphasis on the bottom line, I have to question how much more had been done outside of purely business development for the internet as a technology. Google probably is the one massive trend setter at the moment in terms of coming up with technologies that are actually useful. Search, maps, and the subsearches like video and images have made google look like the prime time in terms of innovation. But the truth is that these ideas are useful. Not everything produced on the internet has been useful. Rather that most things are ecommerce related, blog/bbs related or informational. You may consider security an area of improvement, but I would say that security is more a result of opportunistic businesses taking advantage of horrible programming (e.g. microsoft). The other things that had become somewhat interesting technology-wise are WiFi, Skype and web cameras. The increase in bandwidth allows for better communication than the antiquated phone system. In fact, I don't understand why we still have home phones. It's simply another useless peripheral that occupies more space. We should be employing web cameras over WiFi for maximum efficiency in communications. Of course email is useful, but it's more evidence related to documentating our conversations and for business related notions (like memos). However, the email protocol itself is horrible because it's extremely insecure. Telnet to port the smtp port and you can anonymously send email to anyone from anywhere. But going back to the net, what paradigm shifts might we see? The big media companies have struck down the notion of peer-2-peer software with their avarice. I think the biggest thing is the quality of information that would be made available through search engines. Google obviously is on the case, but most of their stuff deals with keyword searches. With many people publicizing techniques for this avenue, it's become passe. But I like the idea of image searching. The idea should be that an image has properties of Nth quantity. I can see an application where you drop one image in a search bar and all related items come up based on certain criteria you define. Take for instance the image of a building. Things we may want to know about are the location, the material of the building, who designed it, the size of it, the color, etc. These properties should be made searchable in itself. I'm thinking that such an image would be embedded in a special package containing something like XML with descriptions on multiple levels tied to the image. When a search engine goes out to locate this information, it can easily parse through these tags to obtain the information and retrieve it as part of its result set. Problem is getting that information into a large system. Two ways of doing it. One is to build a system that assigns properties when searching for the items in that picture list. Very tough because there are no precedence now known for that type of recognition. So you'd have to build something intelligent enough to recognize those properties. The second way is the easy and cheap way of doing things: have the public do it. Wikipedia, dmoz.org and OSS demonstrate clearly that contribution of information is more efficient building something to collect it. The problem exist in context. We understand the context better than something we can build so it's far better to have society input that information into such a system. Once that system is built with enough information loaded into it, you can build a more intelligent system to massage the data out. Of course the largest issue is copyright. However, what needs to be demonstrated to all businesses is that this system is solely intended for the benefit of mankind as an instrument for aiding people. Think of some of the social benefits. I'm certain the FBI would....

burning dvds on fedora

Thu, 14 Jul 2005 12:36:33 -0600

http://66.102.7.104/search?q=cache:k_14gXv6YNUJ:www.mjmwired.net/resources/redhat8-dvd.html+%22fedora+core+1%22+burning+dvd&hl=en&lr=lang_en|lang_ja

soap clients

Thu, 14 Jul 2005 17:17:22 -0600

http://soapclient.com/XmethodsServices.html

Computer Science Apathy?

Wed, 20 Jul 2005 02:17:19 -0600

Or so says Gates in response to the way US education at the higher level has looked at computer science. Unlike the bubble period, good old comp sci has been falling off the charts, although it's still one of the top majors. The difference is that perhaps the glory of the dot com period's burst students' hopes especially with the declining job market. Likewise, in talking with some people at work with regards to qualified candidates, there just aren't any. Outsourcing seems to be a fad moving away as offshoring's economic benefits bespeak more complexities than it once was purported. So now the US is experiencing another gap in qualified candidates fulfilling solid roles. Of course, the bigger question is whether or not companies are willing to hike pay back up to reasonable levels in order to attract top talent, or just motivate people back into the game. One thing is for certain. Those who stuck it out during the bubble burst and have solid experience are going to be even more valuable than ever. Just recently without having to do a thing, recruiters had been calling my cellphone. I haven't responded because I don't feel the need nor timing is right. Yet signs are pointing in a very positive direction. For someone like myself who has scored with some of the best companies, that makes me look even more valuable in the long term because that gap will continue to grow. Kids coming out of school now with comp sci degrees lack real world experience to back up their theories. However, I have the real world experience of nearly 6 years now (for those recruiters out there, the official date would be around September). Should I stick it out with my company for a good two more years, I'd have a little more than 7 years of experience. That'd put me at a senior level at any normal IT company. But in looking at part of the dilemma, the biggest problem, imo, is the Dilbert/Office Space factor. Office jobs are horrible to start. Computer work is even worse. Cross the two together with the humiliating experience of bad entry level pay, long hours and not-so-glamorous working conditions, it's no wonder why the comp sci field has been avoided within the past few years. The other thing that people might not realize is that outside of hardcore research, the comp sci field is one that you don't have to learn in school to be good at. Some of the best people I've known in the industry never had a comp sci degree. They simply were passionate about what they did, which produced excellent results as programmers. On the other hand, quite a few people in the comp sci field who ended up working were some of the worst programmers I've seen. Perhaps they could take a test in school, but their code would blow chunks. I think the other thing about comp sci is that they emphasize far too much on the theory, which tends to be dry except for the most devout scholars in the field. Most of the theory for the average company is absolutely useless, especially when you get into more mathematical and algorithmic matters. Having the fundamentals behind you certainly helps, but the deeper matters really prove to be something better left in academia. Unfortunately, schools (and students) tend not to differentiate between the computer engineering and IS portion of these degrees. The computer engineering portions are mostly for the IBM styled research labs while IS is what most of us deal with. But again, why torture yourself in school when you can be a humanities major, have a life, a girlfriend and get the same job if you already can teach yourself a few skills?

AJAX Programming

Wed, 24 Aug 2005 22:52:44 -0600

started doing some AJAX programming at home. wow! the speed is unbelievable without the page refresh. at first it was a little confusing because of the asyncronous protocol calls, but once you have a pattern down, it's not so bad. there's some good reusable functions involved. can't wait until the design patterns books comes out for this. let me just say if you haven't started on this, get on the boat. no, i'm not hyping it for your next job, but just as a killer technology to learn.

more AJAX

Fri, 26 Aug 2005 02:32:10 -0600

got my home app finished where i'm now doing updates strictly through my web service controller and javascript. the response time is incredible. almost instantaneous! the search part blitzed me in the response time. it'll be interesting to see what other methodologies and ideas will arise out of this system.

hackers

Sun, 11 Sep 2005 11:07:43 -0600

i do believe that a lot of these incidents at banks and other major websites occur because of insiders that have access to databases and code repositories. there was an incident at AOL where some fired software engineer was being thrown in jail for selling customer information to a spam outfit. the other thing is that you have to really understand the complex infrastructure of a bank firm to know which servers to hit when it comes to obtaining data. there's so many layers to go through before hitting that type of data. also, most companies do a heavy amount of QA'ing before code gets released. so even if a hacker can infect a website, there should only be a brief period where they can do something like obtaining user information. however, what was interesting was seeing this DVDVR message board (using PHPbb apparently) getting hacked. they haven't recovered yet and (considering the technical "expertise" of the people there) i kinda doubt they will. those sites i believe can easily be hacked because the code is open and easily exploitable. plus, most people who set up those sites just know enough to get the basics up. anything more complex, and they ought to be working for a major internet company. now this isn't a rant against open source (trust me; windows has a million more problems than this). but it's interesting seeing something like this get exploited.

Firefox 1.0.7 is out

Tue, 20 Sep 2005 23:34:22 -0600

Get it now. Can't wait until Thunderbird and Firefox 1.5 are released. Should be in the next two months!

fedora core 4

Mon, 03 Oct 2005 11:16:40 -0600

got this baby up and running finally. decided i wouldn't take a chance and bought a new 160 gb hard drive to replace my 80 gb one. core 4 took a little bit of time to setup (there was a weird noise going on with the CPU; maybe my fan is going. so i might just end up getting a new CPU). but the good thing is that the DVD installed fine (as opposed to the utter failure that blew my workstation up at worK!!!). my biggest two concerns this time around was getting the sound to work and making sure my main home directory was mounted properly. the great thing was that neither turned out to be an issue. sound worked just perfectly, as opposed to my fatal experience with core 2 where my sound card (and onboard) was detected but nothing coming out. the USB drive was instantly recognized and even mounted! that was pretty sweet. i was thinking of copying the params so that it's permanently mounted during startup. this gives me total independence and allows me to take my work elsewhere, if say, i got a laptop and needed to travel (of course, i'd need power to hook the drive up, but that's another story). the other thing was making sure my desktop information remained intact. with the exception of a few icons, i haven't had any issues. the real issues was getting yum and apt working properly. the current apt isn't updated with the appropriate repository information so you have to hand edit the file. i encountered a few problems with yum having some conflicts. namely, the old standard library for C++ is needed for things like firefox to be installed, if you do hand installations. apparently, the system has some legacy packages on the disks, but it didn't work for me. however that aside i've been quite happy so far. i still need to experiment with the DVD writer, Japanese IME, gaim, flash, wine, dual monitor system and a few other things. but the basics work and that's important. oh, yeah. forgot to add. the system is WAAAAAYYYY faster!

subversion

Thu, 06 Oct 2005 14:20:15 -0600

using this puppy now for my side stuff. not too shabby. in some ways, it's more intuitive than cvs. for instance using URL style parameters to access files locally or through the net. also, there's a nice plugin for WebDav. i'm still learning more of the features here, but so far i've been pretty happy. some awkwardness for me between cvs and subversion has been the notion of versioning. rather than files getting versioned individually, it seems that the whole project gets versioned. you can commit files and directories individually, but they get recorded as a specific version. the other slightly confusing issue is dealing with the login. it really is just using an htaccess style front end through apache (or something like that) so you just need to access it as if accessing a password protected webpage. still i had some issues with my friend's system on cygwin. gotta try it at home to see if it's an issue with cygwin or an issue of the firewall on his system. i think it might be cygwin, but only one way to tell.

new Dell coming soon!

Mon, 10 Oct 2005 17:15:09 -0600

i broke down and bought a Dell Inspirion notebook. it's got a P4 1.7 ghz, 80 gb hard drive, 1 gb of ram, ATI video card, 14.1" monitor...not bad. i'm considering dual booting it with Fedora Core 4 and Windows XP. I honestly don't need much hard drive space and will probably end up using my external USB drives as my data drives for my user home directories. i already have a similar setup on my desktop. with this, i just create a new user, login through him, attach the drive, then add my primary user back in. and viola! i don't have to worry about copying everything over! the windows side doesn't need much either. again i have two spare usb drives (which i need to toy with) that are 40 gb each. i'm thinking of RAID 0'ing them to give me a spare 80 GB, despite not having any failover. we'll see.... the windows side really will be used for games. i'm eager to test out the wireless connection at my friend's place sometime. then we can get some game action going without me being forced to play from home :)

dual boot for core

Wed, 19 Oct 2005 14:48:47 -0600

http://dis.cs.umass.edu/~msims/dualboot.html

mbr fix

Mon, 31 Oct 2005 15:35:57 -0700

http://www.ntcompatible.com/How_to_remove_GRUB_loader_t28242.html http://www.gnu.org/software/grub/grub-legacy-faq.en.html#q12

dell is dual booting

Sun, 06 Nov 2005 23:32:30 -0700

the OS got seriously fucked over with the MBR taking over. i had to use my friend's windows XP disk to do the recovery since my MBR was only loading an empty Fedora after i blew away the original OS. didn't realize at the time that Dell (idiotically) didn't supply a recovery disk and that you had to do it on your own. yeah part of that was my fault but i'm surprised that Dell would do such a thing especially when almost all over brands seem to supply one. or maybe not anymore. i ended up getting partition magic since i really don't trust some of these freeware utilities. yeah i took one up the ass, but it's pretty reliable and i got boot magic to manage the boot process. i'm still in the process of adding packages. that part is a bitch since some of the stuff i like is old. other than that, it seems to be running fine. next step is to try running my external drive and seeing how that works as my main user drive.

dell laptop

Sun, 27 Nov 2005 01:28:33 -0700

now i'm switched from my desktops to my dell laptop. i did this because i need to pack both systems. fortunately, the transferring of files has been easy on the linux side because, as i've been ranting and raving all along, running my home directory off an external disk has been really beneficial. not only that, but the system is actually faster too! i think my little shuttle box was dying. the CPU sounded sick and probably needed a new fan. on the other hand my laptop has a cool 1.7 mhz non-celeron processor with 1gb of memory. in some areas like the DVD CDRW combo, it's lacking. but the most important thing is that it runs and performs a little better. probably in the future i'll move towards VMWare. but i'm going to wait until the average system utilizes 2 gb of memory (or when memory comes down in price for that). that'll allow me to avoid dual booting.

firefox 1.5 is out!!!

Wed, 30 Nov 2005 22:34:07 -0700

go grab it fast! or just click on the right side and grab the integrated version. should be cool. can't wait until thunderbird 1.5 stable is released.

OpenOffice 2.0 Crashing

Wed, 30 Nov 2005 22:36:38 -0700

i had to switch out my .doc files to .odt files because OpenOffice 2.0 kept crashing, even when i was doing nothing! that really sucks. 1.0 seemed far more stable. i like OpenOffice but seeing it crash randomly makes me hesitant. worse yet, if say you have an excel file open and two other files open, the entire thing goes down. that's a very inelegant handling method.

for my hp printer

Sat, 04 Feb 2006 22:59:54 -0700

http://www.linuxprinting.org/show_printer.cgi?recnum=HP-PhotoSmart_2570

certifying firefox websites

Sun, 12 Feb 2006 00:52:43 -0700

quick though: they should make a certification like Sun or MSCS for websites that are compliant in their CSS and javascript for firefox. it'd good for supporting firefox and user's of the open source community.

Web 2.0

Sat, 25 Feb 2006 05:50:50 -0700

the buzz is huge. tons of new opportunities opening up. technologies springing up faster than you losing money on the stock market. but what is it? http://en.wikipedia.org/wiki/Web_2.0 should we believe the hype? hard to say but there are definitely huge advantages not so much in the paradigmatic definition of it, but rather the technological paradigm that surrounds it. mostly, the outpouring of content that can be aggregated, the services involved in channeling that information, and protocols/front ends (AJAX) supporting them. the two biggest things i see as monstrous potential money makers in this are XML and AJAX. these are enablers to allow distributing, dynamic computing to be alleviated from the strict desktop to a more platform independent vision. some interesting things i've seen are yahoo's AJAX API, web pages being more flexible such as moving content around the page, the lack of refreshes in the browser and the rich layer of CSS that sits on top of HTML, XML, etc. i think two areas which will extensively focused upon in the upcoming years are the improvement of javascript features and the enhancement of CSS. the closer binding between CSS, DOM and Javascript really empowered the web browser since elements within a page can be more readily accessed compared to before. in the future, i see increasing the visual capabilities of CSS such as rounded, angular and customized shapes. right now, for instance, you can use a H1 element with a background color to create a box effect. however, people still are forced to use images partly to generate more stylistic pages. i think CSS can easily address by specifying something like: shape: angular, diagonal-left; also, i can imagine creating more sophisticated color schemes for backgrounds like color blending where you mix hexcodes across percentages to create an effect. or another idea would be to utilize inherent icons that are part of a browser set to create visual images. say an email envelope, a car, a man, etc. in turn, these icons could have various attributes like color, size, etc. ultimately, the idea would be to minimalize the need for photoshop and/or a photoshop type of specialist to handle potentially standardized elements of design. of course, you're not forced to utilize these elements, but for us non-designers, it would be a boon to enable those types of elements through CSS rendering capabilities. with Javascript, i hope they increase the feature/function set. for instance, there's no sort function that i know. sometimes i want to perform a perl-like sort on the browser side, given a complex data structure. part of the problem is that i'd end up creating my own version of a sort rather than utilizing some of the symantics inherent in the language. i know part of what they intend to do is mimic the Java 1.5 upgrade by enhancing the for loops. that's great but i'd still like the syntactic sugar of a perl sort to handle my needs on the client-side.

javascript rocks

Sat, 18 Mar 2006 07:07:36 -0700

well the newer versions do. i'm pretty happy with the event handlers, acccess to DOM, and of course the AJAX related stuff. web applications are just so much more powerful with this stuff. it's funny how a company like google can turn something that was ignored in favor of server oriented languages like PHP, Ruby, Java, Perl, etc. into some more serious outside of form validation. Now, you can do some killer stuff. I haven't added much to this site yet, but give me some time :)

web services are sweet!

Sat, 18 Mar 2006 11:41:09 -0700

i just signed up for a yahoo application id for doing searches using their system the other day. it's pretty slick. you "only" get 5000 queries a day, but essentially, you can get information back from them in a variety of formats including XML, JSON and PHP serialized objects. I tried it using PHP's Simple XML API and web programming has never been easier Also, I tried out Gigablast's web services. The only thing I didn't like about Gigablast's version is that the information is very inconsistent. Part of their information is derived from dmoz. So sometimes you get multiple rows that you may not necessarily want. Then you'd end up cycling through those rows. It's easier if the result set is somewhat static. On the other hand, you might be able to obtain the keywords section from Gigablast's version.

drag and drop yahoo UI stuff

Sat, 01 Apr 2006 23:08:48 -0700

started working with this. let me say that it's pretty damn cool. it does take a little time to get used to in terms of the basic ideas. reading how the event system works along with how DHTML is handled also helps. but so far so good. i'm curious how game play is going to change due to the power of javascript in the browsers. i mean, the thing with ActionScript is that you have to purchase a license to create flash stuff. with javascript, it's native in the browsers. the only hampering element depends on what version of the browser the user has. but that's really not an issue for you as you can force the user to upgrade, much in the same way that games for users to upgrade their hardware to some minimum standards. at any rate, cool stuff. hope to see more.

Livejournal blocks add blockers

Thu, 20 Apr 2006 16:04:00 -0600

well this is pretty lame. it's good though because what this means is that there's a huge market still for people who want to get into the blog software market. i'm a firm believer that you can't always coerce users into making them pay for something they don't want nor need. thanks livejournal for making the barriers to entry even easier!

laptop hard drive died :(

Sun, 30 Apr 2006 11:37:13 -0600

this really stunk. woke up yesterday, messed around a bit on my system then kaput. wasted a day and a half repairing this SOB and even bought a lesser hard drive. the Windows XP part is completely gone. but most importantly i lost all my email data. i did something stupid. i moved the thunderbird folder to a part on my local hard drive rather than copying it. so i lost the whole damn thing. fuck. 2 years of emails just gone like that. :(

killer javascript

Thu, 04 May 2006 13:03:41 -0600

my skills with javascript are getting pretty good. i've learned so much with it that i'm doing pretty advanced things. client-side javascript kicks major ass. i love the tight integration between the browser, DOM, CSS, and accessibility through javascript. one of my next goals is to see how to integrate XSL with it since i'm doing some AJAX application programming. my theory is that all templates should be converted over to XSL. why? portability. it seems as though the transformation parts of AJAX rely too much on putting "view" type of code (as in Model-View-Controller patterns) in the javascript itself. I saw a templating system spring up for Javascript and thought to myself, "Oh no! Not another friggin templating system!!!!" what people should realize is that it's futile creating and using yet another incompatible templating system. there are so many that really make no difference whatsoever. however, XSL means more since both the server and client side do rendering. so imagine having client side components like a sidebar that are initially rendered and cached on the server side. when the client needs to do some replacing, it can retrieving that XSL component and insert it back into place. as a result, this allows one to split the view logic neatly on both the server and client side and provide a nice medium that's reusable. really when it comes to AJAX programming, i think it's preferrable to have a front end controller if you're into the MVC pattern. model layers are data objects serialized into XML. that makes it reusable on both the client and server sides. anyway that's my 2 yen ;)

bought a new IOData TV Capture Device

Sun, 07 May 2006 04:00:04 -0600

USB2. kicks ass. don't know how to plug it into linux but that's okay. the great thing is that it's very portable. even better was that i used my bic camera points and bought the thing at half price! i debated whether or not to pick up a 22" TV or this device and felt that this was definitely worth the trouble. besides, i might later buy a 19" monitor, which can be used in the states as opposed to a regular TV here (well at least from what i know). besides i'd prefer just a larger monitor over anyway.

do-it-yourself

Sun, 21 May 2006 16:15:50 -0600

i think i should stop being my own hardware support person. another fucked up drive. sucks. hopefully, i'll be able to return it later this week.

got new 19" monitor!

Sun, 28 May 2006 05:28:47 -0600

before i had only a 15" monitor for my Gateway. today i got a 19" Acer (XXIDI). not so bad. it's huge! i almost bought the 20" one because it was a reasonable 45500-en. but i hesitated and realized that the 19" was enough (for now :p). now i can watch TV and DVDs and not have them strain my eyes. however, the other benefit is that my 15" monitor was generously donated to my laptop. i had to update the ATI driver though. but now, it works perfectly (er...so far). ah, i never knew what i was missing by having dual head. huh huh huh huh. he said head.

do-it-yourself hard drive

Sun, 28 May 2006 05:41:58 -0600

got it working btw. my fix is in my tech solve section. it was a problem with the jumper settings. that said, i'm tempted to pick up yet another 300 gb hard drive. however, i saw a Western Digitial 320 GB drive today for the same price. mmmmm....20 more gigs :)

bill gates steps down

Fri, 16 Jun 2006 08:53:22 -0600

i don't think any company's chairman stepping down had made such monstrous headlines as this one did. supposedly, bill is just going to be spending most of his time working on the gates/melinda foundation. sounds like he's just going to return to M$ for the yearly shareholders' meeting. without bill and paul, M$ doesn't seem as huge of a threat as in its heyday. ballmer is like george w bush; he's loud, causes issues, but i can't see someone like this as a serious threat except in screwing things up constantly. of course, that could mean M$ might become hungry again. however, the main vision and power behind M$ are pretty much gone. but hey, i can't blame bill for leaving. what more can you do for a company that exclusively focuses on software? having your own foundation with a load of cash behind it means you practically can focus on any other causes you'd like. we'll see what happens.

what really matters in IT

Mon, 19 Jun 2006 09:46:46 -0600

after working for 10 companies with various roles in IT, i've come to realize that the most important thing is not the actual technology itself, but the leadership involved who have visions of how to utilize technology. the current spot i'm at, i'm afraid, lacks any sort of real vision because, as it is implied, the leadership for technology is poor. most of what goes on is merely an enforcement of international standards. more than that though, each group is at war with each other and we end up becoming slaves to the business side because there is no one strong enough to withstand the politics and argue in favor of technology. i've seen companies with the mentality of buying and outsourcing over in-sourcing and utilizing open source. the key with them is that they have some sort of vision for integrating these technologies together. you can use MS technologies and it might not be so bad if you understand how all the components flow together. my suggestion for those that go with MS is to stick with MS. don't try to cross it over with other technologies. keep things similar to avoid waring ideologies and schizoid programmers looking to compete against each other. same with open source. i'm a big open source lover. however, there is a danger of just being an evangelist. you still need capable people of handling it. i've seen places that just rock because the people are top notch. then there are places who use it but just are clueless. why? it's in the fundamentals and vision. in the end, if you lack vision, then your systems become a hodge-podge (as Schwartz likes to put it); kinda like duct tape holding everything together. don't force integration because it increases the cost. instead, use similar technologies that people can become accustomed to easily and make sure the skill set of the people implementing them match up well. don't switch gears in the middle because you'll end up having to support two layers. it's far more expensive and headache inducing than you can imagine.

myspace and online status

Wed, 05 Jul 2006 13:03:45 -0600

myspace has a little online status checker where you can see if a user is online. they did it simply with a little polling trick where they set a timer and continue to send updates to the server. in turn, this allows people to check whether a user is actually online real time or not.

AJAX and MSIE

Sat, 16 Sep 2006 11:03:55 -0600

didn't realize that MSIE does not expose the XSLTProcessor in its javascript like mozilla. that caused a LOT of pain for me. i guess the thing is that Mozilla is just far ahead of the curve when it comes to the features in javascript. i just wish that other browsers close this gap or that more people download firefox or camino to have poor programmers like myself to avoid the pain of dealing with this.

to upgrade or not to upgrade? that is le question...

Mon, 25 Sep 2006 10:36:05 -0600

Fedora 6 is coming up soon in a little over two weeks. but looking at this version i don't see anything imminent that really will make me go out of my way to upgrade to this version.

improvements in bandwidth

Sat, 14 Oct 2006 11:42:28 -0600

it occurred to me that as more applications move online, one of the keys is increasing the bandwidth in networks. currently, outside of servers, the biggest issue for viability is networks. if you examine, for instance, youtube, strictly technically speaking, their biggest bottlenecks were bandwidth usage and possible server implosions. hosting companies need to improve on these areas.

sweetness with mod_rewrite

Tue, 30 Jan 2007 10:51:02 -0700

i've rarely used mod_rewrite but finally came across a situation where I found a great use for it. Say you want one of these so-called "meaningful" urls (i.e. no query string, just a certain url structure). if you use .htaccess you can have the base url (say Review) utilize a handle like php. that way, you don't need to access the server configuration to do something special for your application and not have to append .php (or some other fugly extension to your file/url). but still, what happens if you use css, images, javascript, etc. that depend on a root directory to derive their information? the new url will foobar those elements....unless you rewrite them. using symlinks is painful so instead, you can put a RewriteRule in your .htaccess file (granted your server allows mod_rewrite). so you just need to add a few lines to the .htaccess file: RewriteEngine On RewriteRule ^Review/images/(.*)$ /images/$1 this simply changes the request for your images from /Review/images/some_image.gif to /images/some_image.gif. easy, right?

PHP and interfaces

Wed, 09 May 2007 05:52:28 -0600

I don't know why PHP uses interfaces. It's counterintuitive to me since the signatures and return types for interfaces in PHP have no meaning. You can't specify a return type nor signature value in an interface using PHP so the only thing that you can do is enforce naming conventions and people implementing the function calls. You're really better off coding an abstract class at this point. Interfaces work well, imo, with strict typed languages like Java. Part of the idea would be that a method's return type could be an interface but something like a factory might return the concrete implementation. In PHP, there are no specified return types as part of the method/function call, so you lose the benefit of an interface. If they want to really implement interfaces, they should make it so that return types and parameters for method calls are strictly typed in PHP. Just my PHP rant.

The Network as the Bottleneck

Mon, 17 Sep 2007 18:03:28 -0600

As applications explode become free, hardware performance increases with better CPUs, memory and storage, optimization continues for systems and techniques for scalability improves, the one thing I've seen as a constant is how bad networks are. Certainly, wireless and the penetration of broadband have helped move the internet along, I cannot help but feel that network bandwidth and constraints are really holding back computing to the next few degrees. This notion came to me from the following observations:

A problem recently at work where displaying a 1MB PDF file took around 40-50 seconds with the servers being hosted in the states.
The cost of fiber optic cable between servers to an EMC and the fact that this cost was to be partly mitigated through another internal group's rental for the lines/service.
The problem of ISPs especially the cable companies of throttling users for "excessive usage".
The high cost of networking for bandwidth usage that hosting services and the like charge to businesses.
The various limits imposed by FCC in terms of how wireless is handled.
An entry in Wikipedia describing Web 3.0 as being 1 GB of networking compared with 10-100 mb.
Too many outtages on things like mobiles and connectivity as a result of bad networks.
Someone once telling me in my former securities company how network people were basically the last in line in terms of blame for problems in the application; that is, they had no one else to blame since there were no lower levels that could cause an issue.

If these observations hold true, then I think that there needs to be a renewed interest in pushing the limits of the network. In other words, we need to improve networks in terms of redundancy, cost, speed, security, and flexibility. At the same time, phone companies, ISPs, the cable companies (or few remaining) and mobile servies must find ways to reduce the cost and improve their service, rather than overcharging users for crappy services. I feel that there are too many unnecessary constraints on networks that limit another explosion of technical advances.

More than that, I feel that consumers are simply held hostage by these bottlenecks for no good reason. Obviously, we had past cases like Bell Labs which needed to be broken up at one point. Now, we're facing the same crisis with limited selection of providers (at least in the states) and far too many restrictions (not just on policies for usage). Heck, ironically when I posted this blog, my ADSL device shut down. So I think I'm onto something....