pan-users
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Pan-users] Re: Is this project still alive?


From: Duncan
Subject: [Pan-users] Re: Is this project still alive?
Date: Sat, 21 Jun 2008 01:35:55 +0000 (UTC)
User-agent: Pan/0.132 (Waxed in Black)

Ben Bullock <address@hidden>
posted address@hidden, excerpted below, on  Fri, 20 Jun 2008
23:18:01 +0000:

> On Fri, 20 Jun 2008 05:37:05 +0000, Duncan wrote:
> 
>> JCA <address@hidden> posted
>> a10b0c8a0806190627t6354e362q31562bc8cdeacdfd-JsoAwUIsXosN
> address@hidden, excerpted
>> below, on  Thu, 19 Jun 2008 07:27:50 -0600:
>> 
>>> Version 0.132 came out some 10 months ago. Is this still being
>>> developed?
>> 
>> Well, yes, but not really at the moment.
> 
> Has anyone heard from Charles Kerr about this?

Not that I know of.  He apparently ignores both the user and the devel 
list for long periods of time -- the .nzb file buffer overflow was posted 
to the devel list with no answer, except from us users and some 
distribution pan package maintainers.  On the bug, the GNOME folks bumped 
the bug priority to the highest possible and marked it security, and 
you'd /think/ if /anything/ would get a response /that/ would, but not a 
peep.  Basically the same with the gcc-4.3 patch and the glib 2.16 bugs 
and patches. 

I don't know if anyone has contacted Charles at his personal email (well, 
as personal as we know), but certainly, the bugs should be getting to him 
as the pan developer, and if it's a security bug...

However, again, it fits the pattern.  I don't know of any security bugs, 
but during the 2-3 year hiatus between the last 0.14 C code release and 
the first C++ release (0.90), the lists/groups were entirely silent, but 
for the period of a couple weeks when he did the two additional beta 
releases somewhere in the middle.  Then he went silent again.

Of course, we now know that he was doing the C++ rewrite, but nobody at 
all knew it then.  He had just disappeared, for all intents and 
purposes.  By the two-year mark people were beginning to think it had 
been abandoned.  Not a peep saying otherwise.  No hint at the rewrite, 
which he later said he wanted to keep quite until it was basically 
finished, as he didn't know how long it would take or whether he'd ever 
finish at all and didn't want to get hopes up.  Still, some hint that he 
was still around and working on something special, that he wasn't going 
to talk about at the time, would have been nice...  Even if he'd dropped 
a note to me or one of the other regulars saying "hey, I'm still here, 
working on something I don't want to say much about, but you can at least 
tell the group/list the project isn't abandoned", it would have been /
something/.  But not a peep.

So yes, it /does/ fit the pattern, and 10 months is nothing compared to 
the three years it was the last time, peep or no peep.

I *DO* know that Charles is (or was for a time after he went silent here) 
still active on other GNOME projects.  I forget the name of it ATM since 
I'm normally a KDE person, and all of this is second-hand from a pan user 
that follows GNOME in general, but there's a GNOME torrent app that had 
evidently been going pretty slow that Charles evidently either adopted or 
submitted a whole bunch of patches for (in a classic Charles "great guns" 
phase), that ended up being absorbed and advancing the app quite some 
distance in a relatively short time.  However, even that was probably six 
months ago now, and I've no idea what has happened since then.  Still, we 
can glean from that, that Charles hadn't (at that point) dropped out of 
GNOME developing entirely, but had just stopped doing anything with pan 
for the moment.

>> Really, what Charles needs is a couple other developers to work with
>> him on pan, helping to fill in the dead periods, while still allowing
>> him to come along every couple years and really go to town on things.
> 
> Has Charles expressed an opinion about this at all?

He hasn't, except the very general that he has always made it know that 
he's willing to take patches, as long as they don't kill the GNKSA 
compliance or otherwise screw up pans way of doing things or primary 
goals (including a pretty strict no-bloat, Charles has several times 
trimmed features pretty severely, only returning those that got asked 
for).

During the previous timeout, before which there had been some talk of 
switching to a different database backend, perhaps sqlite, a number of 
devs started experimenting with databases, and Charles did integrate some 
of their ideas into the pan rewrite.  He had mentioned sqlight so most of 
the experiments were with it, but he didn't end up using it and of course 
no one knew he was switching to C++, so he couldn't use the code, just 
the ideas, but he did use those.  

That wasn't a (public) fork just some personal experimentation, but by 
the time Charles surfaced again, there was talk of taking some of the 
work and continuing what was to all appearances a dead and abandoned 
project.  I'm not aware of Charles ever responding to any of that 
discussion at all, altho he certainly should have been aware of it as he 
came back into circulation and read the discussion, some of the ideas of 
which he responded to and ultimately used.

>> However, I've been a regular on the pan lists/groups for years now
>> (since 2002 IIRC) and it hasn't happened yet.  There was a guy (Chris)
>> that helped for awhile, but he was very much either minor, or simply
>> behind the scenes.  Charles was still the primary developer.
> 
> I wonder what's happened to Chris now?

I don't know.  That was pre-rewrite, and AFAIK, Chris was around until 
the long break started.  I saw what might have been one patch from him in 
the rewrite, but nothing else, and I might have been wrong on that, it 
may just have been based on an earlier discussion Charles and Cris had 
had via private email or something.  As far as I can tell, he moved on to 
other projects.

>> I'm not sure why this has been the case, whether Charles is hard to
>> work with, or whether the area simply doesn't interest most developers
>> enough to do more than submit an occasional patch, but that's the way
>> it has been.
> 
> In order to work with Charles, the first step would be getting his
> permission to make modifications. Has anyone actually applied for
> permission to him?

No one has said as much on either the devel list or here.  You can 
imagine how frustrating this is for me, too.  I wish I had the skills...

>> Of course, as most newsgroup folks already realize, NNTP itself is
>> relatively obscure; nothing like web browsers or mail clients, for
>> instance, so maybe it /is/ simply lack of interest.
> 
> Another problem is that one of the big uses of Pan seems to be for
> binary newsgroups. I don't even have access to binary or alt newsgroups
> - it's hard for someone like me to take responsibility for something
> which I can't even test.

Yes, that /would/ be a problem.  However, if it were a development team 
instead of just one person, what I'd like to see and consider the ideal, 
and with, as had been demonstrated by the weekly betas, quite a decent 
number of people at least willing to file bugs and round-trip the debugs, 
I believe it's reasonably workable.  (Some of those bug filers are like 
me, tell me what to put in gdb and I can tell you what I get out, but my 
programming has all been scripting and Basic/VB, the latter before I 
switched to freedomware of course, some actually know what they are 
doing.)

>> I just wish I
>> had the necessary skills to contribute at that level, but I don't, so I
>> simply stick around here and contribute what I can, help on the pan
>> lists/ groups.
> 
> Well, you seem to be able to compile the newsreader from source code,
> apply patches and so on.

Yes.  I guess I'm reasonably functional at a sysadmin level, compiling 
and applying patches, building my own scripts, etc.  Back before I 
switched off of proprietaryware, I had so MS Windows VB programming 
including "advanced" (for VB) Windows API work, and can follow a lot of 
programming discussions reasonably well, knowing the basics of working 
with toolkits and the like.  That has actually a lot to do with the place 
I've found in the FLOSS community, as while I can't actually program, I 
grok the general ideas and can follow discussions well enough to explain 
the basics in easier to follow "plain English".  So I've found a useful 
niche, and on the lists/groups where I'm active, a reasonable amount of 
respect from both devs and users, as a sort of go-between.

Still, it has been personally frustrating that altho I taught myself 
bash, I tried and failed to learn much perl, and have been going to look 
at python but haven't really gotten into it yet.  Would I could do more, 
but I'm in my 40s now, and it really does take more effort to learn 
things than it did as a teen.  I'd love to be a decent coder, but have 
come to realize that realistically, I'll likely never really grok C or C+
+ to the degree necessary to do anything decent, tho I still have a 
chance at the interpreted and managed languages, python, maybe java.  Oh 
well, as I said, I've found a bit of a niche as a go-between, and do my 
best to contribute what I can there.

>> Oh, well.  It is what it is.  Either people with the skills are
>> interested enough to take a major and continuing interest, or not, and
>> it appears not, so Charles continues his mostly solitary developership,
>> and we that lack the coding skills to do more, continue to deal with
>> the on and off pattern, because he's providing the code and we're not.
> 
> I'm not sure whether you're suggesting a fork of the code or not.
> 
> It's quite a big decision to do that, and it would be a very nice thing
> to talk to Charles Kerr beforehand.

Not so much a fork, for all the usual reasons.  Rather, ideally, as I 
mentioned above, a development team.  I've no idea if Charles can 
actually work that way or not, but it's what I'd consider ideal, if it's 
possible.  Otherwise, I don't know.  A team could in theory do better 
than the on-again-off-again pattern we have now, even if no individual is 
half as good as Charles, but as you say, a fork, particularly a no-
permission hostile fork, is a **HUGE** step.

OTOH, Charles has always sort of invited it when asked about changes that 
would threaten GNKSA compliance.  He's always said not on his pan, but 
that if people believe in it strongly enough, they are welcome to fork it 
and do what they want with the fork.  That was certainly in the FLOSS 
general attitude mode, and a real fork, particulary a hostile one, is 
absolutely different than the theory, but he has recognized and invited 
at least in the general case the possibility.  Still, I'm not entirely 
sure it'd be an entirely hostile fork, based on that.  In some ways, I 
think he might even welcome it, as it would give him somewhere to point 
people who wanted features incompatible with his ideas of where pan is 
going.  And I think he'd cooperate and share code, at least until the 
point of divergence to the point it was no longer practical.  So as I 
said, I don't believe it'd be an entirely hostile fork, particularly if 
the work he and those before (the original C pan author was someone else).

There are also alternatives, altho AFAIK pan is the best merged binary/
text client.  On the KDE side, klibido is quite good as a binary-
harvester type client, no posting and no text, but boy, it really goes to 
town harvesting binaries!  Until it came along, pan was far and away the 
best we had on *ix for binaries, altho there were various text clients, 
some of which sort of handle binaries as well.  With klibido, pan isn't 
quite as critical for binaries any more, and as I said, there are other 
decent text clients, so pan itself isn't as irreplaceable as it was for 
awhile.

Perhaps, if Charles isn't amenable to team work or forking, it'd be 
better to investigate the alternatives.  Still, pan's original dream was 
as the "Pimp-Ass Newsreader", aka PAN, and particularly with the C++ 
rewrite and the automated server handling it offers, pan is certainly a 
good way toward that goal.  Thus, it'd be a shame to drop it or to 

....

Taking a look at pan's home, contact page, it seems Chris is still listed 
with a rebelbase address, Charles is listed, and Matt, I /think/ the guy 
who owns the rebelbase domain name, is listed.  There's of course the 
general pan@ alias as well.

In pan's about box there's another person listed in the credits for the 
thread decoding.

If one were seriously considering helping out with pan, or just wanted an 
official word and/or an opinion from the others mentioned as to where 
it's going and/or about the practicalities of working with Charles, those 
contacts would be the place to go next.  You might also post an inquiry 
to the devel list, to perhaps catch the folks who did the database 
experiments and the like, and anyone else who might have info and be 
reading there but not here.

You may also wish to contact the GNOME folks and see what they might 
suggest, particularly given the security bug and inaction on it to this 
point.  Certainly, they must have some interest in at least having 
someone that would do security and preferably gcc/glib type update 
patches.  Again, preferably not to fork, but at the minimum, someone who 
could work with Charles and have commit privs for security and toolkit/
toolchain updates.  One would think that's pretty critical, and it's not 
being addressed at this point, so at least getting it addressed should be 
in the interest of everyone in the community, including Charles.

-- 
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman





reply via email to

[Prev in Thread] Current Thread [Next in Thread]