Ionrock Dot Org

I was thinking about PHP frameworks today and how they are becoming closer to their Python/Ruby relatives. Both Python and Ruby have real HTTP servers (think Apache) that are often used for development. Python has developed more robust solutions through WSGI, Zope and CherryPy. CherryPy specifically is an excellent server in that it is mature, fast, stable and well coded (some of the devs are my friends, but that friendship was spawned after seeing the power of CherryPy so it is still not too biased ;) ). I think Ruby also has some similar implementations outside the Webrick server included with the language.

PHP is different in that it essentially was originally just an extension of Apache. Apache would process a .php with the PHP Apache module and that was that. A "framework" was usually just a set of files you included and the command line tools you see in Rails or with Paste were no where to be found. This was not a bad thing at all, but it seems as Rails has gained traction the command line version of PHP has been brought to the forefront along with CGI interfaces to PHP so that the deployment is closer to that of Rails/Paste/Python apps. The frameworks also seem to have jumped on this ship by providing similar command line tools and working closer with the central PHP libraries.

I think the good side of this is that PHP is getting great frameworks that help to get the productivity gains you see from Rails. The cost though seems to be the easy $10 hosting packages and simple integration. This cost though seems totally appropriate as you see more VPS hosting and virtual servers. This also falls in line with the idea of a home server, which it seems both Microsoft and Apple are beginning to embrace. And wouldn't you know, but my current employer is also embracing this concept of everyone can have servers.

There are many reasons for this movement that involve many years of dealing with specific deployment issues. The idea is essentially capitalizing on increases in computing power, but more importantly it is acknowledging the skewed ratio of time spent developing versus the time spent in integration and deployment. Agile methodologies for development espouse constant integration and virtualization simple attacks the issue from a different angle.

I for one am a fan. The big unknown is maintenance of servers that are potentially locked down or run based on images is concerning in that there is a strange issue that pops up when you want the latest and greatest while still keeping apps running. This is the problem I have found with frameworks in general. They make a new release with new features along while breaking a good portion of your application. When you start talking about hundreds of these apps on a virtual server, upgrading core libraries that are used becomes a scary proposition and the result is you end up seeing crazy sandboxing. This will probably lead to a new wave of management applications meant to be the IDEs for keeping track of what applications you have running. It will essentially be another bisecting of the problem that will bring up new problems.

I suppose the moral of the whole story is there will always be a need for software developers!

Last night was our showcase and things went pretty well. We borrowed equipment which is always hard for us. As a group, we come from a rather punk background, but when it comes to sharing equipment and everything we really are babies. Of course, the guys from Aloke have always been insanely awesome when it comes to helping us out, so it is easy to let our silly worries drift away in their kindness.

It was a great night and we were able to see good friends rock. The best part was generally seeing everyone's improvement. I am not saying any band was bad in the past, but rather it was fun to see the great get greater. Aloke added a second guitarist, which added to their power. They also have come off of a great recording with Albini, so they were totally on their game. The Subjects also brought along the rock. The vocals were loud and awesome. They also rock the drums something crazy. I'm a huge drummer fan. Speaking of drums, The End of the World had Stephan, the lead singer, beating down the rhythm. This really was a cool image and sound as it brought out different feelings in the songs I hadn't heard from them before.

A nice dance party wrapped things up and all were happy and spent. Thanks!

Last night I was a little tired of having a few facebook requests sitting in my inbox and decided to log in. Two of the requests were from some random folks I didn't know. Being used to the MySpace way of life, I was going to approve them as they seemed harmless. Facebook promptly informed it was dumb to add them as friends if I didn't even know them. I took the advice and let them both go.

My Facebook experience is matching up very well with my Friendster and Orkut experiences. None provide anything useful to me at all an only require me to work. I just cannot stand this model. YouTube is another social site that asks for content but the results could be fame, backing up a video, or sharing a video with friends. I think I have a flickr site laying around somewhere because I wanted to do the exact same thing at one time so this does make more sense to me Facebook/Friendster/Orkut.

If you notice I did not include MySpace simply because as a musician is does provide value. It gives everyone an opportunity to hear our music. That is really it. I can't tell you how many shows we have played with someone new who says, "I am excited to see you guys, I heard your stuff on MySpace".

I suppose in my mind, a site much provide value outside of the network. Facebook has seemed popular because it is the first "social networking" site with a serious API and way to make thirdparty apps. From a developer point of view this does seem exciting, but it still doesn't provide intrinsic value to the users.

My theory is that depending on the network for all the value makes it impossible to create something like google. If all you have is the number of users to attract and provide value then all you really have is a fad. The critical mass is expensive technically as well as removing value from users. People enjoy network based sites because it provides relationships, but when you have crazy amounts of people it becomes strained. Imagine going to a bar the size of the Rose Bowl? Sure, they could give funny hats or something to help find people, but that point the concept or metaphor is broken. People go to a bar to meet people *and* drink. Even if no one is there, you can still grab a beer and watch TV.

For the record I don't believe adding value is easy. Adding value without adding tons of work is a hard problem. But, searching massive amounts of text and indexing the web was also hard to solve and that seemed to pay off, so it makes a little more sense.