Weekly Edition
Return to the Front page
| |||||||||||||
A day in the life of emacs
Your editor is often asked that most fundamental of Linux-user questions:
vi or emacs? The answer - that both editors often come in useful over the
course of a working day - tends to please nobody. The truth of the matter,
however, is that most of the serious work of producing LWN is done in
emacs. Until very recently, the current version of GNU emacs was 21.3,
which was released on March 19, 2003 - almost exactly two years ago.
Your editor got to wondering about the current state of emacs, and whether
it was still under active development or no. Some time digging through the
emacs development mailing list turned up a few interesting things.
First and foremost, it should be said that the emacs developers are, indeed, active. Whenever the project gets around to making a new release, emacs users will be surprised at how much as been done - more on that shortly. It was surprising to see that Richard Stallman, the creator of GNU emacs, remains very active in its development. He may not produce as much code as he used to, but he is active in the discussions, and still functions very much as the final decision maker on patches. When RMS makes a decree, things happen that way. A reading of Richard's postings indicate a real concern for the utility of emacs and the creation of a useful user interface. Emacs detractors may differ, but the fact is that quite a bit of thought is going into how emacs works. Development is not the only issue to be found on a list like this, of course. Back in December Ben Wing requested permission to use parts of the GNU emacs manual in the XEmacs manual. This sort of reuse would seem to be just the sort of freedom that the GNU project is working for; XEmacs is free software, and its manual is licensed under the GPL. Unfortunately, since the GNU emacs manual is licensed under the GFDL, it is not possible to reuse portions of it in the XEmacs manual. Mr. Stallman's responses indicate that he has no problem with this state of affairs:
I did not choose this license with a view to its effects on you; it
is the general FSF policy for manuals. However, the fact that it
is inconvenient for XEmacs does not strike me as a disadvantage.
After all, you have been uncooperative towards us for 10 years, and
you don't see that as a disadvantage. We don't owe you anything,
not even small favors.
The XEmacs developers would appear to have gone away empty-handed. Shortly thereafter, Steve Youngs showed up with an announcement of a brand new emacs fork called SXEmacs. It appears to be a new version of XEmacs with different coding conventions, Windows support removed, and various other changes planned. Not much discussion resulted, but Mr. Youngs is still working on SXEmacs. At the end of January, Per Abrahamsen proposed that emacs go into a "regression fixes only" freeze so that a release could actually happen. Nobody even responded. On February 7, Richard Stallman noted that he had rushed out version 21.4, which adds a single security fix to 21.3. This move surprised a number of developers who had been telling people about the great new features 21.4 would have. Richard suggested instead that the next release should be version 22, since "It has plenty of new features." A plan to use negative version numbers for test releases (e.g. 22.1.-998) was, fortunately, turned down.
There are hundreds of other changes; the NEWS file has all the detail anybody could want. As for when emacs users will see all these changes: it's hard to say. Mr. Stallman has never been willing to project release dates for software. In this case, back in December, all he would commit to was: "It isn't around the corner, but I hope we are getting closer to it." (Log in to post comments)
A day in the life of emacs Posted Mar 3, 2005 2:14 UTC (Thu) by dlang (✭ supporter ✭, #313) [Link] anyone else find irony in RMS being adament about requiring the GFDL (prohibiting it's use by other projects) and Debian deciding that anything uder this license needs to be removed?
GFDL does not... Posted Mar 3, 2005 12:03 UTC (Thu) by hummassa (subscriber, #307) [Link] prohibit the use of the docs by other projects. But, it *is* incompatible with otherwise-licensed documentation and, as such, cannot be merged with one of those. IMHO, RMS is right IRT the XEmacs project.
GFDL does not... Posted Mar 3, 2005 17:51 UTC (Thu) by TxtEdMacs (guest, #5983) [Link] I guess this is why I will continue to use xemacs and I will never sully the name of Linux with a "GNU" prefix.
Freedom to RMS is more akin to his being dictator and definer of the work "freedom" and in this case not even being a benign dictator.
I still hold RMS in high regard for the essence of his work creating the FSF, but his disdain for all other avenues toward the same goal grates on me. [I am NOT including the "Open Source Software" as one of the parallel movements.]
GFDL does not... Posted Mar 4, 2005 5:34 UTC (Fri) by komarek (guest, #7295) [Link] That Richard wants to maintain control over work for which he has the copyright, is not akin to being a dictator in general. The FSF policy is GFDL, and that is what covers the GNU Emacs manual.
That the fellow from XEmacs asked RMS to "reconsider" his enforcement or selection of this license is interesting, and gutsy. But the tone of that fellow's post was "maybe you can compromise, and hence mend some fences". This seems a bit unlikely:
1) RMS doesn't compromise on his projects. He decides. He seems to care little or nothing for convenience, and principal is everything.
2) Attempts to mend fences in the past have never made any headway. Unless the XEmacs team has something to offer (perhaps soliciting the approval of some XEmacs copyright holders, in order to adjust the licensing), then XEmacs has nothing.
3) And let's not forget: GNU Emacs was the first Emacs. Derivitive works of Emacs that hid their modifications are *the motivation* for the creation of the GPL. RMS discovered that not everyone believed in the golden rule, and some who did would still sign NDA's that prevented them from following it. After being taken advantage of, RMS found a way to defend himself and his work: copyleft. XEmacs is one of these derivatives, one with a long history.
As I understand it, licensing issues prevent XEmacs from contributing improvements back to Emacs. That XEmacs is open at all is nice, but that doesn't help the mother of XEmacs (which is Emacs, before GNU Emacs, iirc). RMS is upset that anyone would ever violate the golden rule, and treat their neighbors poorly. Maybe that's life, but that doesn't mean RMS has to accept the status quo.
I really don't understand the dictator argument. It seems similar to those arguments that compare Red Hat to Microsoft. RMS is not trying to take over the world. He isn't trying to take over anything. He is trying to defend his work, and the work entrusted to the FSF. He also tries to help those who share his priorities and morals (through FSF legal support, for example).
The GNU/Linux debacle is unfortunate. I wish RMS would have said "I will call it GNU/Linux, because it appears to me as the GNU system plus Linus Torvald's excellent kernel." But he didn't. And doesn't. In trying to make his point about the GNU project remaining relevant, and in trying to keep the name of GNU alive where GNU software is used, he has said some strange things. But his actions have *never* been dictatorial, as far as I recall. And his actions are worth far more than his words.
-Paul Komarek
Principals Posted Mar 4, 2005 19:35 UTC (Fri) by Ross (subscriber, #4065) [Link] I'm fine with someone sticking to their principals above convenience. Infact, that's one of the best attributes of RMS.
But I don't see how _not_ "sharing the software" is really part of his
And to the best of my knowledge XEmacs is under the GPL just like Emacs.
Principals Posted Mar 4, 2005 20:13 UTC (Fri) by komarek (guest, #7295) [Link] I see that you are right about the GPL, and I agree then that the problem is probably copyright assignment. Now that it's not so late at night, I've taken the time to read http://www.xemacs.org/About/XEmacsVsGNUemacs.html, and this verifies the copyright assignment problem.
Back to the original reason I posted, sp_ware's comment that RMS wants to be a dictator still seems wrong and ignorant. That he set up the FSF and GNU projects, and gave up his copyrights to the FSF, suggests strongly that he does not want to be a dictator. Any disdain he shows for competing Free software could easily come from disliking the waste of effort of duplication (that xemacs link above quotes him on this, w/r/t XEmacs).
Thank-you for pointing out my mistake.
-Paul Komarek
GFDL does not... Posted Mar 4, 2005 20:45 UTC (Fri) by TxtEdMacs (guest, #5983) [Link] quote: "... Richard wants to maintain control over [his] work ..."
I have NO problem with that, however, RMS has problems with other people making control decisions with software (and perhaps even documentation) which he was neither the creator nor the maintainer.
That smacks of a dictatorial temperament, regardless of the fact of whether he lacks the power of enforcement of his decrees.
I have instances in mind, but too many details can lead to a nasty tenor so it is better to cease here.
GFDL does not... Posted Mar 4, 2005 0:16 UTC (Fri) by dvdeug (subscriber, #10998) [Link] Somehow, the XEmacs people have a different story of the history between the two projects. Given that RMS is refusing to cooperate now, I tend to look askance on any claim that all failures to cooperate in the past have been XEmacs's fault.
GFDL does not... Posted Mar 5, 2005 18:46 UTC (Sat) by giraffedata (subscriber, #1954) [Link] When one assigns blame for a failure to cooperate, it's all about perception of fairness. RMS clearly believes that, over the past 10 years, FSF has contributed enough to the cooperation (the Emacs code and license, I presume) that it's only fair for Xemacs folks to contribute copyright assignments. Apparently he also believes that such copyright assignments would be fair exchange for the documentation license Xemacs is requesting now. So refusing the documentation license isn't hypocritical; it's all part of the same failure to cooperate, and the fact that RMS isn't willing to contribute even more to the cooperation doesn't mean Xemacs is any less at fault for the lack of cooperation.
GFDL does not... Posted Mar 7, 2005 6:19 UTC (Mon) by dvdeug (subscriber, #10998) [Link] Xemacs pulled out because RMS wanted to be lead dictator in Emacs. Efforts at reconciliation have failed because RMS wants the successor to be 100% compatible with Emacs, even though it's not possible to be 100% compatible with both and XEmacs is better designed in some of those places. Copyright assignment is a pain in the ass, but there's no evidence that if FSF asked for something specific from XEmacs, like XEmacs people have done here, that they wouldn't assign the rights. The FSF certainly has the right to use the code without an assignment. Honestly, the FSF has provided code under the GPL, and so have the XEmacs people; why are the XEmacs people not cooperating because they aren't copyright assigning their work to the FSF, when the reverse isn't happening? the fact that RMS isn't willing to contribute even more to the cooperation doesn't mean Xemacs is any less at fault for the lack of cooperation. I find the concept that it's all XEmacs fault for not cooperating to be nonsense. The XEmacs people have a number of complaints about when RMS wouldn't cooperate.
GFDL does not... Posted Mar 9, 2005 11:42 UTC (Wed) by jschrod (subscriber, #1646) [Link] Those of you who want to read the ``other point of view'' to this biased opinion that it's all XEmacs' fault: Please have a look at http://www.xemacs.org/About/XEmacsVsGNUemacs.html. The author tries hard to give credit to both sides, much to the disdain of jwz.
Speaking of him, if you want to read an additional (very strongly biased) view of the Emacs/XEmacs chism, you can also check out Jamie Zawinsky's Web Site http://www.jwz.org/doc/lemacs.html where you can even read the emails of RMS and of Richard Gabriel. IMHO it clearly shows that it's both sides faults.
Joachim
GFDL does not... Posted Mar 10, 2005 7:13 UTC (Thu) by giraffedata (subscriber, #1954) [Link] This is the second comment that has referred to a concept that "it's all Xemacs' fault," but I don't find that that position actually shows up anywhere in the article or the comments or the article's references.All I see is a statement by RMS that Xemacs refused to cooperate. As I pointed out above, that's a nonsense statement -- cooperation isn't something that one party does. So he's probably talking about fairness. But even then, it's a leap from believing the Xemacs people are being unfair to believing that they are 100% at fault for there being two versions of Emacs (assuming that's a fault at all).
GFDL does not... Posted Mar 10, 2005 8:21 UTC (Thu) by jschrod (subscriber, #1646) [Link] You wrote: ``the fact that RMS isn't willing to contribute even more to the cooperation doesn't mean Xemacs is any less at fault for the lack of cooperation.'' (emphasis by me). This was the reaction to the point of dvdeug who wrote ``I tend to look askance on any claim that all failures to cooperate in the past have been XEmacs's fault.'' With your statement, you made a contradiction to his scepsis and thus supported the concept that the chism is XEmacs's fault.``any less at fault'' as an answer to ``all failures ... have been XEmacs's fault'' is quite clearly a position, isn't it? And I reacted exactly to that position. Joachim
GFDL does not... Posted Mar 10, 2005 21:23 UTC (Thu) by giraffedata (subscriber, #1954) [Link] "not any less at fault" is fundamentally different from "completely at fault" or even "somewhat at fault." In fact, I have reserved judgment on who, if anyone is at fault and how much.dvdeug said something else besides that he looks askance at this claim by RMS; he gave the reason -- that RMS is today doing the same thing he accuses Xemacs of doing in the past. All I said is that the reason is wrong -- RMS's refusal to give the documentation license doesn't absolve Xemacs of any fault it might have. dvdeug then added some different reasons to point the finger away from Xemacs and at RMS.
RMS is *very* active Posted Mar 3, 2005 2:28 UTC (Thu) by tnoo (subscriber, #20427) [Link] As a user of CVS Emacs I stumbled over several bugs / inconveniences. RMSwas always incredibly fast (~hours) responding with a working bug fix for his pet-project. And this also applies to some other developers. I found this really impressive. Tnoo
A day in the life of emacs Posted Mar 3, 2005 4:36 UTC (Thu) by dhess (guest, #7827) [Link] Does the CVS version of emacs support attaching a gnuclient/emacsclient from a tty to a gnuserver/emacsserver that was started by an X11 emacs?
This feature is useful if you have, say, a gnus session running at work on your X11 workstation, and you want to attach to it to read email from home over an ssh connection without forwarding X11.
The current release of emacs doesn't support this feature. I dug around on the mailing lists and saw something from one of the emacs developers to the tune of, "We can't do it because of the design of the emacs event loop."
XEmacs does support this feature, and it's the main reason I use XEmacs.
A day in the life of emacs Posted Mar 3, 2005 8:04 UTC (Thu) by dann (guest, #11621) [Link] No, it does not, but the multi-tty branch does.See: http://lorentey.hu/project/emacs.html.en That branch that tracks CVS head very closely supports multiple terminal and X11 frames at the same time, even terminal frames on terminals supporting only 8 colors at the same time with terminal frames on 256 color terminals. It works greak.
Vi vs. Emacs? Posted Mar 3, 2005 5:54 UTC (Thu) by ncm (subscriber, #165) [Link] I don't know why anyone would see Emacs and Vi as opposed. I run Vi from within Emacs, in Viper mode ("M-x viper-mode"). Detailed editing is a lot less stressful and more efficient that way, but I still have easy access to all the Emacs bloat^Wfeatures.Essentially, in Viper mode Vi's insert mode is generic Emacs. The Vi command mode is like regular Vi command mode, but regular Emacs commands still work. There's practically no conflict. Emacs always lacked Vi's "dot" command, a serious omission; now it has it. IMHO viper-mode should be the default; then Vi would be just another subset of Emacs.
Vi vs. Emacs? Posted Mar 3, 2005 17:30 UTC (Thu) by mwh (guest, #582) [Link] Bet vi still starts up faster though!
(Mandatory flameproofing: I am not being entirely serious here)
Vi vs. Emacs? Posted Mar 5, 2005 13:33 UTC (Sat) by aglet (guest, #1334) [Link] The fact that X?Emacs doesn't come with the very useful "dot" command is a bit annoying, however, you can get one: see for example http://www.esperi.demon.co.uk/nix/xemacs/site-wide/vi-dot...
Vi vs. Emacs? Posted Mar 5, 2005 18:30 UTC (Sat) by giraffedata (subscriber, #1954) [Link] I don't know why anyone would see Emacs and Vi as opposed.I think it's because so many people, including me, believe they cannot efficently maintain expertise in both. When I first started using Unix, I knew I couldn't afford to learn and keep up with more than one editor, so I looked around and picked the one that most closely met my needs. Incidentally, it's the same reason I try not to use any program that has its own editor built in. One set of keystrokes is all I can handle.
Vi vs. Emacs? Posted Mar 13, 2005 15:10 UTC (Sun) by anton (guest, #25547) [Link] >[...] many people, including me, believe they cannot efficently>maintain expertise in both. When I first started using Unix, I knew >I couldn't afford to learn and keep up with more than one editor >[...]
Well, expertise. I am an Emacs fan and have used it almost
As for learning, when I started with Emacs, I had had two months of
CLIPBOARD vs PRIMARY Posted Mar 3, 2005 10:43 UTC (Thu) by tialaramex (subscriber, #21167) [Link] Does anyone (the LWN editor included) know whether GNU Emacs has correct behaviour concerning the X clipboard in this next version?
XEmacs correctly distinguishes between the CLIPBOARD and the PRIMARY selection, but GNU Emacs seems, iirc, to ignore CLIPBOARD altogether, making it difficult to use in conjunction with modern apps that obey ICCCM and Freedesktop standards. I'm told that yet another M-blah-blah-blah option was added, but that's no substitute for simply doing the Right Thing™.
CLIPBOARD vs PRIMARY Posted Mar 3, 2005 11:30 UTC (Thu) by cyd (guest, #4153) [Link] Short answer: yes, but it's not on by default.
You can make killing and yanking consult the clipboard if you set x-select-enable-clipboard to t. The default is nil, except on MS Windows where there is no primary selection.
Another emacs relative Posted Mar 3, 2005 12:02 UTC (Thu) by rmstar (guest, #3672) [Link] Have you looked at climacs? I find it quite interesting. It is not yet fully functional, but it is under active developement.
Another emacs relative Posted Mar 3, 2005 17:09 UTC (Thu) by ncm (subscriber, #165) [Link] Is this the one started by the noted net.kook Erik Naggum?
Another emacs relative Posted Mar 4, 2005 4:46 UTC (Fri) by flewellyn (subscriber, #5047) [Link] Naggum is known for being outspoken, but he's hardly a kook. I found his writings on various aspects of language design and computer use to be quite fascinating and insightful.
Another emacs relative Posted Mar 4, 2005 9:17 UTC (Fri) by larsbrinkhoff (guest, #4638) [Link] No, it's not started by Erik Naggum.
A day in the life of emacs Posted Mar 6, 2005 20:45 UTC (Sun) by zonker (subscriber, #7867) [Link] How about a day in the life of Vim?
A day in the life of emacs Posted Mar 10, 2005 10:24 UTC (Thu) by leandro (guest, #1460) [Link] Where are my Debian emacs-snapshot packages?
Seriously, they've been coming for quite a while now...
Free software needs free documentation Posted Mar 10, 2005 10:28 UTC (Thu) by forthy (guest, #1525) [Link] I'd like to remind RMS on that subject. Apparently he has forgotten why. Free Software is about sharing software, and being able to help. If I have some ideas how to improve Emacs, and other people want them, too, I must have the right to change Emacs, and when RMS is not agreed with that, I'll have to fork. That's what Richard Grabriel had to. If I'm forking a project, and creating incompatible extensions (the main reason for forking, isn't it?), then I also have to fork the documentation. Therefore, following the rule that free software needs free documentation, the documentation must allow forking the same way as the code. If it's not possible, the documentation license is not free. If the FSF requires copyright assignments to take the forked code back, it's their fault. The copyright assignments are there to protect the code with FSF's power, not because it's legally necessary. So, FSF (and RMS): If someone refuses to assign the copyright for some personal reasons, it's not your business to complain. You still can use the code. I'm using GFDL for a few documents, too. I use a limited version of the GFDL, without the non-free stuff - no invariant sections - (though I have the FSF's back cover text, which is mandatory policy for my official FSF project. I think it's foolish, because it's inaccurate as soon as someone forks the project). My suggestion to the GFDL issue is that if you take material, you either have to take it *with* all the invariable sections and cover/back cover texts, or *completely without*, and use a new cover title. That way, the variable part would still be free. Otherwise, it would be quite stupid. Cover: "This is the Emacs documentation" by Richard M. Stallman Backcover (mandatory texts): "The FSF prints this manual"
Free software needs free documentation Posted Mar 13, 2005 15:27 UTC (Sun) by anton (guest, #25547) [Link] >Therefore, following the rule that free software needs free>documentation, the documentation must allow forking the same way as >the code.
The GFDL does. That's not the problem (as far as I understand it).
Linux will probably have a similar problem with GPLv2-only vs. GPLv3
|
|||||||||||||
![[Emacs screenshot]](/images/ns/emacs-22-sm.png)