Dan Widyono on 10 Jan 2008 11:56:09 -0800 |
Hi Brian, > At the same time, I must also point out two other things: > (1) I don't think it's 100% fair to blame the said flaw on the registry. I agree COMPLETELY with that statement. In fact, I could care less about the registry in this discussion. What I meant in my reaction to your bringing up the registry was: using it as an example of how or why syntax needs to improve, merely hinders your argument, IMHO. I'm simply trying to help hone in on what would move this conversation forward. I feel bad for the effervescing eclipsed equine, you know. One problem I have with this entire conversation is I haven't read any shining example of a good configuration API that has consistently improved syntax across multiple diverse vendor products (on one OS, much less across multiple OSes). Even though Windows provides a Registry, it really doesn't do much of anything with respect to what Matthew wants; rather, it just enforces typing in a very useless way because there is no DTD, no spec, no good explanation (AFAIK; I'm not a Windows programmer, but a quick google search did not inspire confidence in finding much) of what to do with the types. As an example, there ought to be a REG_DATE instead of a REG_SZ with a date inside it; that should be explicitly defined and explained to programmers. Perhaps an ODf for configuration files. *That* is what really is necessary in order to start achieving what Matthew wants. The API should then arise organically from that standard specification (or worked on in parallel, but the API should follow the spec). So, long story made short: to discuss syntax consistency, focus on specs, not e.g. the horrid Registry implementation which does nothing to improve the situation (reminds me of MS OOXML, but that's a whole other can of worms). > (2) The overall message that I was trying to get across was that, though > it is commonly misused, at lease Windows and OSX provide *some sort* of > framework for doing the things that Matthew is describing. I agree. In Linux there is no standard library framework for such work. That could enable useful further interactions, and that's speaking as a manager (policy driven), as a sysadmin (consistency, reproducability, and maintainability), and as a user (ooh look, turn this dial and that happens!). I assume programmers would love having another library to play with. :) Cheers, Dan W. ___________________________________________________________________________ Philadelphia Linux Users Group -- http://www.phillylinux.org Announcements - http://lists.phillylinux.org/mailman/listinfo/plug-announce General Discussion -- http://lists.phillylinux.org/mailman/listinfo/plug
|
|