Why We Continue to Add Features to Slash 0.3/0.4
In the previous installment of this series, we talked about the reasons why CTDATA built and maintains its own Slash implementation. In that article, we pointed out that our goal is to converge our code with the main Slashcode distribution once the Bender version is finally released.
This article approaches our situation from the opposite direction: If CTDATA does plan to merge its Slash enhancements with Bender, why do we continue to add new functionality to Slash 0.4, like enhanced search capabilities and RSS support?
Needs of Existing Customers
Throughout this Weblog, we've alluded to the fact that we have several working instances of Slash 0.4, including a large corporate client who uses our code as the core of a knowledge management system for a departmental Web Site. A few of these instances continue to run Slash 0.4 because there are conversion issues which make it very difficult to transition them to Slash 1.x or Bender.These issues include:
- no institutional support for Apache / mod_perl
- no insitutional support for mySQL, or a preference for an enterprise SQL database, like Sybase or Oracle
Linux Transition Issues
Until we began working with Slashcode 1.x a few weeks ago, CTDATA had no Linux experience whatsoever. Our experience with Slash to that point had been on Solaris and (believe it or not) the Windows platforms.Therefore our internal issues were:
- no experience operating Linux as a bastian host
- no spare server-class hardware on which to install Linux
Increased Traffic on Our Sites
In spite of the fact that we are not running the "latest and greatest" version of Slashcode, traffic on Web Sites we manage has grown steadily since we began to deploy Slash 0.4.Most of our users have no idea what version of Slash we are running. However, those that have seen Slashdot before recognize the similarity and often expect us to have Slash 1.x features whether we are running that code base or not. This is the reason that we have rolled out RSS Slashbox support at this time.
If we can figure out a cross platform way to make home page personalization (i.e. cheesyportal) scale within the Slash 0.4 architecture, we will probably develop and deploy that feature as well.
Feel free to let us know if you have any comments, questions, or suggestions. Send an email to Dave Aiello or leave a comment attached to this story.