Randy McMurchy wrote: > TheOldFellow wrote these words on 07/30/05 00:49 CST: > > >>For after 6.1 - Has anyone thought about ways to measure usage of BLFS >>packages - I'm not talking about popt, but courier vs postfix vs >>sendmail or Gnome vs Kde vs a-small-and-beautiful-wm etc? >> >>LFS has the 'registered user' thing on the website - perhaps we could do >>something similar, but of the major packages. Then we can drop things >>that are (a) not used by any editor and (b) only used by people who >>could do the build without help. > I didn't mean to start a discussion now, while we are concentrating on a release, but...
> Something like this would be difficult, if not impossible. Trying > to gather subjective information like you suggest is at the whim of > folks being cooperative. We cannot count on that. Nothing worth doing is ever easy. ;-) I remember when BLFS was said to be 'too hard'. > Richard, if a vote was taken and KDE outweighed GNOME, would that > be cause to remove GNOME from the book? > > I don't think so. Nor me, but it would help in deciding how much effort to put into each one. On this point, a key factor is 'is there an editor using the package?' as well as 'is there a user for the package'. I suspect that there will always be some editor who likes these overstuffed 'humungous' desktops. :-) > Recently, Bruce asked if XFree should be removed. A couple of guys > responded no. This, I hope, would be enough to keep it in the book. I think it depends on how good a job we can do. If there is no editor actually using it, then how can we know that the build is good enough for use. Just 'cos it compiles an' installs 'aint enuf. > My reasons for suggesting that Courier be removed is due to the fact > that the BLFS version of Courier is grossly behind and there seems to > be no Editor interest in keeping it up. The BLFS instructions for > Courier are no longer accurate for the last 7 revisions of the package. > This can only lead to a support headache if we release a BLFS version > with the current version of Courier. A really good reason, if I may say so. > Can you think, or identify, just *one* other package that is similar > to Courier in this respect? Nop. > To summarize, I don't think there is any way to accurately measure who > is using what packages, other than submit a survey. And a survey > would surely be skewed due to user preference of those who decide to > contribute to the survey. It is unrealistic to think that we could > get an accurate measure of usage. We could put 'spyware phone-home' patches in the book. (Joke!) More seriously, if users were aware that their favourite web browser was going to be dropped if they didn't register, then they might. > It is up to Editors to decide what should be in BLFS and what should > not be in BLFS. as it should be, of course. My only point with Courier is that there is no > Editor that seems to be interested in keeping up with the package. > I cannot think of one other package (Exim came to mind, but Bruce > recently updated it) that isn't updated on a timely basis. Does Bruce use Exim? I wonder how he knows the build works otherwise. Now, this isn't a personal criticism, so don't take it that way. I'm just thinking about the methodology, now that I've agreed to use it. I need to answer these questions for myself. Richard. -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
