Opensource server software will ultimately be better than "enterprise" options

I'm sure it's not worth nearly what we pay for it, but for all my complaints it's quite a capable system.

I'm not that familiar with autosys yet, so I went to look for documentation from the vendor.

That was mistake number one. The Computer associates website is confusing to start with. Many clicks later I was able to find the link to access the pdf manuals for autosys but I had to register. So register I did, but the link that was emailed to me to verify my registration didn't work in firefox! Luckily (for sufficiently small values of luck) I am on a windows machine so I could open the link in MS internet explorer and validate my account. The whole process was futile however, as CA only allow you to download the documentation for the most recent version. This site is running a major version behind, so the manuals that were there were no use.

I managed to track down the relevant version of manual by searching for 'autosys' on our internal wiki - a victory for collaborative distribution of knowledge over the company who should be trying to help me 'leverage our investment in job scheduling infrastructure', whatever that means.

The manual, however, is woeful. It's badly written, verbose in the fluff and disturbingly lacking in detail. I have been unable to find much of the information I need and will have to resort to poking around the database (which thankfully is Sybase not proprietory to the product). If that doesn't answer my questions I'm kinda stuck. I can't read the source code (obviously) and the (non-authorised) mailing lists haven't had any traffic in the one week I've been monitoring them.

Speaking of Sybase, it is also hard to get good documentation on that as well, but let's not get into that here.

The problem is that job scheduling is incredibly boring. Why would CA bother to soend time and money producing good documentation since the people who actually use it don't make the purchasing decisions.

Compare this to a (hypothetical) open source job scheduling project. The people who write the code would be actual users of the system. It will be valuable for them to document it becuase that will help them and their co-workers. A mailing list would have developers and users swapping problems and solutions. If the original authors got bored or went out of business or whatever there would be no problem because the other users can continue the development (or hire someone to do it if they don't have the skills). For the licensing and management costs of these sorts of "enterprise" software products, it would be very viable to hire a developer or two to keep an opensource alternative humming. And when your boss wants you to implement some sort of job schedule not supported, you can add that feature and fold it back in for everyone else.

I know that this is all obvious stuff, and most people in the opensource community knows that it is true, but it is still often thought that the major benefit of opensource is cost and that if you've got money to burn you are probably better off buying a commercial product that will come with support and documentation (and someone to blame).

My experience is that exactly the opposite is true. Now if I can just get some budget to implement an open source job sceheduling tool...