So, how about that switch to Subversion, eh?
I'd very much like to rename lib/modules/_Image.pmod to Image.pmod and src/modules/Image to _Image, to get rid of image_magic_index, and with svn I could do that without messing everyones working copy up...
I'm in favour, especially if the right people are dependent enough on the online cvs browser to hack up one covering the svn repository on fairly short notice (a week or month, for instance) after (or before) the switch. I noticed I wasn't, last time I started. (And that's even worse these days.)
Still in favour though, regardless of whether that happens or not. :-)
I'm in favor, but why not just use something like FishEye? It's free for open-source use, is reasonably attractive, and has lots of features that Code Librarian doesn't have.
Bill
On Jul 9, 2006, at 11:45 AM, Peter Bortas @ Pike developers forum wrote:
I'm for, but we need to fix Code Librarian.
It should be noted that the Code Librarian deployed at pike.ida.liu.se is now where near what the HEAD of the development version is. The question really is if it is too complex for such a small project as Pike.
FishEye is listed as something CodeLibrarian should be evaluated against on my source-tools TODO, but last time I checked it was missing some changelog and branch-features. I still don't know if we could actually run FishEye as I'm stopped by the evaluation application from downloading it.
I'll have a look at downloading it again.
I've taken another look now and downloaded a copy to take a look at it. It has some drawbacks:
1. It's all Java and Regexps. We'll need a new modern computer with some horsepower to run this. Not a showstopper per see.
2. It's running Java Server Pages in it's own webserver. I have nothing but bad experiences with JSP in Tomcat/Apache, but it is concievable that it might work better in their solution as it is all Java. Showstopper if I'm going to administer it.
3. There is no source. We can't change it unless we decompile it. Showstopper for me.
4. I won't agree to the NDA. It's not to horrid, but it forbids reverse engineering. See 3. Showstopper for me.
Since I won't agree to the license I can't evaluate it any further.
(I apologize in advance if this seems overly snippy, but I somewhat expected the response given, so I feel I need to vent a little bit.)
I've been using the application (FishEye) for a little over a year now; both at home, and at work. I think some of your points seem less than credible, or at least indicate criteria designed to produce a preselected outcome. To rebut:
1. I've run Fisheye on an ultra 10 without any significant hardware issue. Granted, it uses a moderate amount of memory (as does any java application, or roxen, for that matter), but I don't think that should be a show stopper.
2. There isn't really anything to administer... it runs in its own container, and you simply start/stop it. The repository at work is 90 million lines of code with 100+ committers; I've never had to touch it.
3. I'm not exactly sure why this should be a show stopper before you've even looked at the product. There are plenty of projects using it that you can get a feel for the product. More specifically, what exactly would you think you'd want to change?
4. see point 3.
I guess I don't understand what the reluctance to even evaluating it is. I mean, does it really make sense to be spending time reinventing the wheel, just so that it's written in pike? Certainly the limited resources at hand could be better spent on more worthwhile endeavours. Fisheye certainly has worthwhile features (tarball generation, searching, rss feeds) that I think merit more than the very curt dismissal you've given it.
Bill
On Jul 9, 2006, at 3:05 PM, Peter Bortas @ Pike developers forum wrote:
I've taken another look now and downloaded a copy to take a look at it. It has some drawbacks:
It's all Java and Regexps. We'll need a new modern computer with some horsepower to run this. Not a showstopper per see.
It's running Java Server Pages in it's own webserver. I have nothing but bad experiences with JSP in Tomcat/Apache, but it is concievable that it might work better in their solution as it is all Java. Showstopper if I'm going to administer it.
There is no source. We can't change it unless we decompile it. Showstopper for me.
I won't agree to the NDA. It's not to horrid, but it forbids reverse engineering. See 3. Showstopper for me.
Since I won't agree to the license I can't evaluate it any further.
Of course it was the expected answer. I'm a bastard and will happily complain about any solution, let's keep to the technical points though.
1. I explicitly said it wasn't a show stopper. But we will need to upgrade our hardware.
2. If there wasn't anything to administer it would magicly run on pike.ida.liu.se right now, without needing disk, CPU, backup, access to the repository or user management.
3&4 How will evaluating it give me power to change the code when in fact the NDA forbids med to change the code?
I _have_ evaluated it as far as I can. Anything further would mean agreeing to a licens I bloody well won't agree to for no good reason. I'm also not perticularly keen on reinventing the weel which is why I went through 22 3rd party solutions (open and closed source) last year to see if I could find anything that would work. The best out there meeting most of the requirements is Trac (http://trac.edgewall.org/). The reason we are not currently running it is that it's a bitch to compile on Solaris. I've done it once to test it, but it needs a lot of baby-sitting each time it needs updating.
And, as has been mentioned, the current codelibrarian (as used on as an example Opera Softare) does support most features we would want.
And adding more is trivial.
On Sun, Jul 09, 2006 at 08:39:18PM -0400, H. William Welliver III wrote:
- I'm not exactly sure why this should be a show stopper before
you've even looked at the product. There are plenty of projects using it that you can get a feel for the product. More specifically, what exactly would you think you'd want to change?
we need to be sure that we can use this for time to come. no entity can guarantee that they will be around and supporting their software for any length of time (not for free anyways) who is to guarantee that the owners of the product don't suddenly turn around and want to charge money for it?
therefore the only assurance that remains is access to the source (and the right to change it.) the risks with not doing so are simply unacceptable for a project like this. (yes, this includes the use of roxen platform on the pike site, but there i believe/hope is a special contract between ida and roxen)
I mean, does it really make sense to be spending time reinventing the wheel, just so that it's written in pike?
but at this point CodeLibrarian is already written. so it is a bit late for that argument. and yes, it does make sense to reinvent the wheel if the new wheel is easier to maintain or the old wheel is not free software.
Fisheye certainly has worthwhile features that I think merit more than the very curt dismissal you've given it.
i can only support the dismissal. i can't even find the license terms without registering. i find such behaviour deceptive and not worth my time.
greetings, martin.
we need to be sure that we can use this for time to come. no entity can guarantee that they will be around and supporting their software for any length of time (not for free anyways) who is to guarantee that the owners of the product don't suddenly turn around and want to charge money for it?
Sigh. I don't want to get into a open vs. closed source argument here. In this specific case, the makers of Fisheye will grant a free, perpetual license for open-source projects.
therefore the only assurance that remains is access to the source (and the right to change it.) the risks with not doing so are simply unacceptable for a project like this. (yes, this includes the use of roxen platform on the pike site, but there i believe/hope is a special contract between ida and roxen)
you're making statements based on a particular ideological viewpoint, and that's fine, but with the exception of access to the source code, the license is not really any different than a free software license. and really, I think it's a bit dramatic to say that the risks are unacceptable. it took me a total of 15 minutes to install and configure the software. If for some reason in the future i was prevented from using the software, i'd only be out 15 minutes. It doesn't mean my [or the pike repository's] data will be trapped.
I mean, does it really make sense to be spending time reinventing the wheel, just so that it's written in pike?
but at this point CodeLibrarian is already written. so it is a bit late for that argument.
but it doesn't support svn, nor a number of other features that products on the evaluation list do include, so that argument is still perfectly valid.
and yes, it does make sense to reinvent the wheel if the new wheel is easier to maintain or the old wheel is not free software.
It hasn't been established that one "wheel" or another is anything, and that's my point. They're running those boxes with plenty of other non-free software, so arguing against using software that isn't free- as-in-speech to solve a particular niche problem just doesn't hold water for me.
Fisheye certainly has worthwhile features that I think merit more than the very curt dismissal you've given it.
i can only support the dismissal. i can't even find the license terms without registering. i find such behaviour deceptive and not worth my time.
I realize that you're on a strict free-software only diet, but I think in this case, you're being alarmist. I'm trying to be pragmatic here; there's a piece of software that I suggest will make everyone's life easier (and I truly believe that). It's not free-as-in-speech, though, and I think it's silly-beyond-words to not at least look at it.
BTW, the license is here (nothing special, it's a standard closed- source license):
http://www.cenqua.com/fisheye/license.html
greetings, martin.
On Sun, Jul 09, 2006 at 10:19:38PM -0400, H. William Welliver III wrote:
Sigh. I don't want to get into a open vs. closed source argument here. In this specific case, the makers of Fisheye will grant a free, perpetual license for open-source projects.
where do you read that? (it's not in the license page)
therefore the only assurance that remains is access to the source (and the right to change it.)
you're making statements based on a particular ideological viewpoint,
no, i am making them based on experience with non-free software.
and that's fine, but with the exception of access to the source code, the license is not really any different than a free software license.
huh? that's like saying with the exception of having protection, a helmet is no different than a wooly hat.
having access to the source (and more) the crucial differentiator i care about in the license.
and really, I think it's a bit dramatic to say that the risks are unacceptable.
they are to me. i could not be doing the work i am doing now, had i accepted to work with non-free software (or worse, create non-free software)
it took me a total of 15 minutes to install and configure the software. If for some reason in the future i was prevented from using the software, i'd only be out 15 minutes. It doesn't mean my [or the pike repository's] data will be trapped.
you must not forget the time it takes to learn to USE the software, the investment of teaching users, answering questions, and all that. sorry, 15 minutes just don't cut it. the longer you use something, the harder is is to change to something else. you are grossly oversimplifying things here.
They're running those boxes with plenty of other non-free software, so arguing against using software that isn't free- as-in-speech to solve a particular niche problem just doesn't hold water for me.
two wrongs don't make a right. just because there is already non-free software in use, is no excuse to continue that. the decision wether to use a particular product should be based on its own merits (that includes the license) and not based on what other unrelated software that might be in use.
It's not free-as-in-speech, though, and I think it's silly-beyond-words to not at least look at it.
it is not silly. you might be prevented from creating competing software if you look to closely. the existance of an NDA suggests as much (anyone got a link to that?) i prefer to play it safe and stay away from such things unless i really need to look.
BTW, the license is here http://www.cenqua.com/fisheye/license.html
ah, thanks.
In exchange for the License(s), the Licensee shall pay to CENQUA a one-time, up front, non-refundable license fee. At the sole discretion of CENQUA this fee will be waived for non-commercial projects.
pike is not a non-commercial project. owned by the university it is now an academic project, but that does not exclude commercial exploitation by the university.
The Maintenance Period commences on the day the license is issued and continues for twelve months thereafter unless otherwise stated in writing by CENQUA. The Maintenance period can be extended for an additional fee.
this reads to me that the license is granted yearly and needs to be extended every time (even if it's free for non-commercials)
this opens exactly the can of worms i am afraid of, and is far from an unlimited license that would be needed.
The Licensee may not assign or otherwise transfer the Software to any third party.
if pike changes hands again, then what?
greetings, martin.
where do you read that? (it's not in the license page)
It's on their website. It's how I got a free license for use with my open-source projects.
huh? that's like saying with the exception of having protection, a helmet is no different than a wooly hat.
having access to the source (and more) the crucial differentiator i care about in the license.
no, that's an invalid comparison.... you can have a gpl license revoked if you violate the license terms... in that way it's no different, and that was a key concern you pointed out.
and really, I think it's a bit dramatic to say that the risks are unacceptable.
they are to me. i could not be doing the work i am doing now, had i accepted to work with non-free software (or worse, create non-free software)
What does that have to do with CVS browsing software?
it took me a total of 15 minutes to install and configure the software. If for some reason in the future i was prevented from using the software, i'd only be out 15 minutes. It doesn't mean my [or the pike repository's] data will be trapped.
you must not forget the time it takes to learn to USE the software, the investment of teaching users, answering questions, and all that. sorry, 15 minutes just don't cut it. the longer you use something, the harder is is to change to something else. you are grossly oversimplifying things here.
I have a hard time taking this argument seriously. I mean, do you need to spend hours trying to figure out every website you've ever been to? do you really think the key audience of code librarian would not be able to make the transition from code librarian to viewsvn or fisheye? I mean, it's not rocket science...
They're running those boxes with plenty of other non-free software, so arguing against using software that isn't free- as-in-speech to solve a particular niche problem just doesn't hold water for me.
two wrongs don't make a right. just because there is already non-free software in use, is no excuse to continue that. the decision wether to use a particular product should be based on its own merits (that includes the license) and not based on what other unrelated software that might be in use.
No, but your tone suggests that you have an ideological agenda here. Your suggestion that the "wrong" license should automatically disqualify any piece of software seems just as closed minded.
It's not free-as-in-speech, though, and I think it's silly-beyond-words to not at least look at it.
it is not silly. you might be prevented from creating competing software if you look to closely. the existance of an NDA suggests as much (anyone got a link to that?) i prefer to play it safe and stay away from such things unless i really need to look.
I think peter mispoke. there is no nda to use the product. you just can't reverse engineer it. there's no downside to trying it. who knows, it might actually fill the needs, or give someone an idea of how to do something better in code librarian.
BTW, the license is here http://www.cenqua.com/fisheye/license.html
ah, thanks.
In exchange for the License(s), the Licensee shall pay to CENQUA a one-time, up front, non-refundable license fee. At the sole discretion of CENQUA this fee will be waived for non-commercial projects.
pike is not a non-commercial project. owned by the university it is now an academic project, but that does not exclude commercial exploitation by the university.
the software is open source, and the vendor will provide a non- expiring license for pike.
The Maintenance Period commences on the day the license is issued and continues for twelve months thereafter unless otherwise stated in writing by CENQUA. The Maintenance period can be extended for an additional fee.
well, that's for commercial support. the license itself doesn't expire.
this reads to me that the license is granted yearly and needs to be extended every time (even if it's free for non-commercials)
this opens exactly the can of worms i am afraid of, and is far from an unlimited license that would be needed.
The Licensee may not assign or otherwise transfer the Software to any third party.
if pike changes hands again, then what?
ask them for a new license, perhaps?
you know, you could just ask them if you were interested in actually understanding how the licensing worked.
On Sun, Jul 09, 2006 at 11:11:03PM -0400, H. William Welliver III wrote:
having access to the source (and more) the crucial differentiator i care about in the license.
no, that's an invalid comparison.... you can have a gpl license revoked if you violate the license terms... in that way it's no different, and that was a key concern you pointed out.
you are misinterpreting things. i am talking about revoking the license at the whim of the owner, not about a condition that i know to exist to begin with.
and really, I think it's a bit dramatic to say that the risks are unacceptable.
they are to me. i could not be doing the work i am doing now, had i accepted to work with non-free software (or worse, create non-free software)
What does that have to do with CVS browsing software?
i'd be one of the people using the software. and i am voicing my opinion on the issue. unless you are saying that i should not do that, i find the actual reason for the software irrelevant. what i said applies to any software being used.
I have a hard time taking this argument seriously. I mean, do you need to spend hours trying to figure out every website you've ever been to? do you really think the key audience of code librarian would not be able to make the transition from code librarian to viewsvn or fisheye? I mean, it's not rocket science...
i did not say that, all i am saying is that you can't just ignore that time needed for transition completely. it is not 0. getting used to a new system always takes time, and having a change in the interface does slow you down. (you may have bookmarks that you'd need to recreate, etc)
No, but your tone suggests that you have an ideological agenda here.
of course i have. i want to preserve my freedom to always be able to reuse any piece of software i come in contact with, and not be frustrated by the fact that i can't fix a problem just for legal reasons.
Your suggestion that the "wrong" license should automatically disqualify any piece of software seems just as closed minded.
closed minded is only if you have not experienced the other side. i have experienced the frustration that non-free software brings with it (for me). i am not disqualifying the license automaticly, but after consideration because of bad experience.
I think peter mispoke. there is no nda to use the product.
ok, we can drop all the arguments surounding that then.
The Licensee may not assign or otherwise transfer the Software to any third party. if pike changes hands again, then what?
ask them for a new license, perhaps?
that only works, iff the company is still around then, and still interested in giving out new licenses. that is what i meant in the first paragraph in my first mail of this thread.
you know, you could just ask them if you were interested in actually understanding how the licensing worked.
i am here trying to explain why i am not interrested...
greetings, martin.
Why buy in to a tool you're not at liberty to hack to fit your needs?
I can see how I'd do it if I didn't value my needs much, or if I had a very good feeling about being able to have somebody else fix my needs for me, say, as was the case between Linux and BitKeeper, but failing both, which I'd say Pike does, I don't see why the license isn't a show stopper.
Hm. Except if I'd firmly believe that I wouldn't ever have any needs that aren't already fixed by someone else by the time I get them. I have never ever had that happen to me in online resources, though.
If it didn't fit your needs (by whatever means available), why would you "buy" into it? I don't see how the license changes that.
However, if it met your needs without having to hack into it, what do you gain by not being able to hack into it? I don't lose sleep at night because I can't change it... to be honest, the thought never crossed my mind. If it didn't meet my needs (or couldn't be made to), it wouldn't have mattered if it was free, I wouldn't be using it.
Truth be told, I've only infrequently not had my needs taken care of before me. Most of the time, they involved pike. I guess that makes my needs less than cutting edge (apparently with the exception of pike).
Don't get me wrong, I'm not trying to be a shill for the Fisheye people. I'm just trying to suggest a pragmatic approach to looking at the options. We're talking about what is essentially an ancillary piece of software, and I don't see the harm in setting up a demo for people to look at and play with. Perhaps that means my software licensing morals are looser than everyone else's here (I happily use closed software when it makes sense and works for me), but that's certainly something I can live with. To me, using Fisheye, even in a transitional way, to get to an SVN repository seemed like a no- brainer. If that's not the case, that's fine, and this will be the last you hear out of me about it.
Bill
On Jul 9, 2006, at 11:00 PM, Johan Sundström (Achtung Liebe!) @ Pike (-) developers forum wrote:
Why buy in to a tool you're not at liberty to hack to fit your needs?
I can see how I'd do it if I didn't value my needs much, or if I had a very good feeling about being able to have somebody else fix my needs for me, say, as was the case between Linux and BitKeeper, but failing both, which I'd say Pike does, I don't see why the license isn't a show stopper.
Hm. Except if I'd firmly believe that I wouldn't ever have any needs that aren't already fixed by someone else by the time I get them. I have never ever had that happen to me in online resources, though.
On Sun, Jul 09, 2006 at 11:42:46PM -0400, H.William Welliver III wrote:
However, if it met your needs without having to hack into it, what do you gain by not being able to hack into it?
^^^^ (i suppose that's loose)
you can decide that in one moment? your needs also never change?
i am with peter here, i have never seen any piece of software (free or non-free) that does everything perfectly. i always develop ideas for new features, or find things that could be improved. and then it is good to have that option, even if it is not used, because at least it will prevent me from whining about it.
I don't lose sleep at night because I can't change it...
i do. at least as soon as i run into a problem i would like to fix.
greetings, martin.
pike-devel@lists.lysator.liu.se
-
H. William Welliver III -
H.William Welliver III
-
Johan Sundstr�m (Achtung Liebe!) ＠ Pike (-) developers forum -
Marcus Comstedt (ACROSS) (Hail Ilpalazzo!) ＠ Pike (-) developers forum
-
Martin Baehr -
Martin Bähr
-
Martin Nilsson (Opera Mini - AFK!) ＠ Pike (-) developers forum
-
Martin Stjernholm, Roxen IS ＠ Pike developers forum
-
Per Hedbor () ＠ Pike (-) developers forum
-
Peter Bortas ＠ Pike developers forum