The Dawn of Slashdot
Dave Aiello wrote, "The other day, I was searching Google for references to Slashdot. I was looking for information that I did not know, whether it was new information or old."
"What I found was an article about Slashdot that appeared in LinuxWorld in 1998. The article, written by Rob Malda, is described as a 'technical history' of Slashdot. I found this article really interesting because it provides insight into the evolutionary changes in Slash engine code base."
Read on for more analysis of the article...
Dave Aiello continued:
The article also gives some insight into Rob Malda's technical experience. For instance, the choice of mySQL was not entirely based on Rob's personal preference for OpenSource software products (at least initially). Instead, it was governed by issues of cost and database server availability on the DEC Alpha / Linux platform:
I was quite familiar with SQL after using Microsoft's and Sybase's respective servers to write assorted applications. It seemed like the ideal way to black box my data storage. Neither of these servers were good options though, as each would have cost significantly more than I could even dream of paying.I spent some time with three different freeware possibilities. mSQL, PostgreSQL, and MySQL. There were pros and cons to each, but eventually I settled on MySQL. It lacks some of the nice features other major databases offered. The biggies were:
- Stored procedures
- Subselects
- Foreign key support
{After reading the article, Rob Malda said that he never really considered using Microsoft SQL Server and Sybase because he "had terrible luck with both of them, from Sybase's
clumsy administration, to MS SQLs bloat and tendency to crash. I
hated them both with a passion. I only mentioned them to illustrate
that I knew that SQL was the way to go."}
Nevertheless, a lot of these points justify the development of Bender, also known as Slashcode 2.0. This is the version of Slashcode that will support several different SQL database alternatives. Why shouldn't you be able to run a Slash-based site on top of an Enterprise-class SQL database?
Coming back to the fact that Rob used Microsoft and Sybase in his pre-Slashdot career, it's interesting to note some additional technical drawbacks to using them as the basis for a Slashdot-like system. These two database servers are historically worse at managing blocks of text than is Oracle. Also, I ended the database choice quote before Rob talks about the performance advantage mySQL has when running simple queries.
Rob also has some interesting things to say about the choice of scripting language for Slash. He said:
As far as I was concerned, the two real contenders were PHP and Perl. All I knew was that either would be much better than the IDC (Internet Database Connector) and ASP (Active Server Pages) I had written so much of during my nine-to-five job.
This is interesting because you have to think back to what each of these tools was like in 1998 (and before) in order to appreciate what he is saying. I worked on an ASP-based project at a major data processing company in 1998, and I know from experience that Perl would have been a much more palatable choice then. The only problem I had was that I did not know how to write CGI scripts back then -- I only knew how to write batch jobs that were executed from cron or the command line.
In summary, I find this article quite interesting. It puts Slash 0.3 and earlier distributions into proper historical context. It also indicates that the early development of Slash was driven on purely practical, rather than ideological, grounds.