The boring life of Jerod Poore, Crazymeds' Chief Citizen Medical Expert.

Case of the Vanishing Medication Pages

The latest problem: bits and pieces of .shtml pages are not being displayed. E.g. any of the Keppra pages other than Keppra's basic information page. If you look at that page you'll see the short disclaimer at the top along with the Google search bar and links to all the other pages about Keppra. Toward the bottom of the page the links are repeated along with shameless begging for support, links to other sections on the site, and the bigass disclaimer. If you click on a link to any of the other Keppra pages (e.g. Keppra's efficacy) you won't see any of that.

The missing pieces are separate files. When I was actively updating the medication pages (and if and when I do so again) I was converting pages from .html to .shtml so any changes to things like the disclaimers, page navigation, etc. meant updating one small file instead of many large files. The files are one the server, they just aren't being displayed.

For those unfamiliar with the HTML code, it looks like this:
< !--#include file="somefile.htm" -- >

The problem is the default for Apache (the software that deals with the web) on the server we're on is to not allow SHiTty Markup Language includes, what it calls server side includes. It's supposed to be a simple override to change that. But no matter where and how and how many times I try to ram that override down Apache's throat it refuses to respond.

In this world of open source there are so many possible combinations of operating system components, and versions of those components, it's next to impossible to determine just where the problem is. Is it a CentOS 4.2 thing? An Apache 2.0 thing? A Plex 8.6 thing? My money is on Plex, because Plex makes everything weird. Case in point, I set up a job to run hourly. It runs, but it doesn't work as I expect it. Meanwhile a similar job runs daily and does exactly what I want done hourly. Except I can't find where in the daily schedule it's run so I can give the hourly job the same parameters. The cron logs give me no clue as I can't even find a reference to the daily cron job as having been run at all! It's Plex! It's magic!

All this really, really makes me miss the Big Iron days when 90% of my fellow programmers never heard of Usenet, Fidonet, or ARPANET until the Morris worm made it into the headlines. By the time ARPANET evolved into the Internet most of them forgot about it all. Besides, who wanted to deal with domain addresses that were Or email addresses that looked like frodo!unauthorized! Back then operating systems were all proprietary and nobody in their right minds altered a single bit. The applications, mostly accounting software, were quasi-open source. The businesses bought them, but the source code usually came with it so the software could be modified. Or it was all developed in house. Back in the 1980s and '90s most places ran a mix of homegrown software and modified software they purchased to run their business and account for everything. Almost everyone who owned the same model of computer was within a release, at the most two, of the operating system as everyone else. Hardly anyone's accounting package looked like anyone else's, even when that software came from the same vendor.

These days it's the complete opposite. If more than one person is involved I doubt if two offices of the same company in the same town have anything close to the same combination of *nix, SQL, and whatever other OS components they need, running on otherwise identical hardware. But the enterprise software they run looks and works exactly the same else in the world (allowing for language differences) that same software is running. These days there's a debate going on about vendors supplying their source code and how the open source movement is changing the way developers think, and all the intellectual property issues and so for.


One thing that didn't change is utility software, although its relevance did. In the Big Iron Age most of us had a lot of the same utility software that made managing various aspects of system administration easier. Why? Because we read the same magazines and typed them in. I couldn't wait for them to catch up and get a way to transfer the source code to a PC so I could upload it. For awhile one of said magazines had the code digitized into something that looked like a cross between a bar code and Rorschach spat. Somehow scanning that with an optical reader and dealing with the reader's software was less complicated than a BBS, a listserv, or even something on the nascent AOL and similar services.

With all the different *nix dialects and variations of basic functions (e-mail, ftp, etc.) I've come to learn that a piece of utility software some random person wrote to make their life easier has become a de facto piece of the OS. To the point where it's just assumed one has it, to the point where further development of the OS is predicated on that software being around, but it isn't included in the standard distribution of the OS.

Some time in the next decade or the one after that they are so going to laugh at the idiots who thought open source operating systems could work for anyone other than people who were nostalgic for the 1970s, when everyone had to build their own PCs from scratch and tailor the operating systems for each box, or for people with a strong DIY ethic who want to do the same thing. But for businesses? For ISPs? It's madness.

OSOS. Oh shit, oh shit.

No comments: