|
From: <ds...@ch...> - 2005-09-28 21:24:40
|
I wrote this the other day, then put it in my draft folder... (So, I'd g= o with the "tracker page" concept.) > Timoth=E9e Lecomte wrote: > > I was wondering if it wasn't the good moment to do a fresh release of > > gnuplot, I mean "gnuplot 4.1". >=20 > I think it's way too early for that. Worse yet, I'd have to be the one= =20 > to do it, and I don't see where I could steal the time for that right=20 > now. Last time took about us 6 weeks of preparation, the last two of=20 > them nearly full-time on my end of things. Yes, but that was a major upgrade spanning back 5+ years, having switched= archive locations to SourceForge, a fully new home page by Ethan--many p= roblems that would've been spread out had releases been done more often. = My thinking is the code has gotten much cleaaner and the developers with= CVS access have been very prompt about jumping on bugs that find their w= ay into the code. > People have got used to a > "release early --- release often" strategy. Sorry, but with the=20 > licensing issues as they are, that's a non-option for gnuplot. >=20 > That set aside, I don't think the code currently has the level of=20 > maturity people rightfully expect from a release version of gnuplot.=20 > There's way too much on-going activity on the 'adding features' front=20 > for that. We have to let the dust settle a bit before a release can be= =20 > considered in earnest. Agreed. But still, I think compared to the position we were in when deve= lopers decided to release 4.0, we're already much further along. >=20 > The most we could usefully do right now is a general review of open=20 > patches and feature requests, with a view towards making a three-way=20 > decision for each and every one of them: >=20 > 1) outdated ---> tag as such, and close. > 2) keeper --> tag 'release-critical', integrate in good time, then clos= e > 3) later --> tag 'later', keep open Also, how about on the web page we have a "future items" list. Something= we all have accesss to and won't get lost. That is, rather than writing= the "new features" subsection of the release notice _after_ the release = is ready, let's do it _before_ the release process is begun. That will h= alt the new features added to the release candidate for a few months. So I will offer up a couple things for the list if we were to start one: Image Plotting: We've covered the major output formats except for Window= s, I think... I know little about windows resources. Binary Data Input: I think Petr and I have shaken out the bugs and preve= nted any kind of major program breakage that could come from trying to in= terpret binary data. I understand this may be a bit dodgy for some becau= se of the low use for interactive users. But it is aimed for interchangi= ng data between programs, not trying to read someone's generic, small, un= defined raw data. Faster Data Input?: On that note, after the fact of using binary data to= speed up data transfer for images came a discussion about the scanf() fu= nction having major slowdown for long line reads... something I only unde= rstood in concept, i.e., having to rescan previous entries on the line or= something. But of course, this probably contributed to the problem of t= ransferring ASCII image files taking many seconds. Will this scanf() slo= wdown be addressed in this release? Or are we at the mercy of any compil= ers on this one? Passing an X window ID into the program: I don't think this is in yet. = I'm not particularly motivated to include it, but someone else requested = that I add it. It seemed such a small addition that it should be bug fre= e, or easy to fix if a bug does come up. (Writing the xWidgets program t= o demo the thing was by far more difficult.) More advanced key placement: Needs debugging. No crash bugs, instead th= e "that shouldn't go there for this option" kind of thing. I'll get to t= his within a month when I get back to my home base. Some compile problems with #ifdef chunks in gnuplot_x11 based upon config= urations at the ./config command line. Again, within a month. Raise/lower for X windows: Don't know the status of that one. May be so= me group reluctance because other window platforms might not have it. Co= de wise, no brainer. Dan |