Can we directly contribute yet?

18 views
Skip to first unread message

T. Goodchild

unread,
Apr 16, 2012, 1:47:17 PM4/16/12
to EXPath
So…Florent is the only person with repository access. First of all,
I’d like to say that Florent is much smarter than I am, and will
probably always be my technically superior to me in every way. I
think the EXpath and CXAN projects are exactly what the XML world
needs, and we have Florent to thank for both of them. :)

However, I have some serious concerns about the future of these
projects.

• The EXPath website is now over 3 years old (launched 2009-03-29).

• On the EXPath “Specs” page [ http://www.expath.org/specs ], of the 8
modules listed only half (4) have a release date sometime in the year
of 2011.

• The EXPath wiki similar shows no “last updated” times later than
2010.

• The EXPath wiki still has no means to register for a username. A
**wiki**—a system invented for collaboration and cooperative
editing.

• The EXPath wiki also says that registration may be accomplished with
an e-mail request to Florent. In January of 2011, I did exactly
that. However, Florent said that I should instead send him an e-mail
with the changes I’d like to make, so that he would make them for me.
(1) It’s frustrating to be told “send an e-mail to register” and then
hear “Actually, I’m not giving anyone access.” (2) What is the point
of a wiki at all if only one person can contribute?

I mean no disrespect, but honestly—this is the most *closed* “open
source” project I have ever seen. It doesn’t matter how awesome the
vision of EXPath (or CXAN) is if it’s only allowed to grow at the pace
of a single person’s spare time.

These projects are the seeds of something wonderful. But they were
planted in 2009, and they need more than one caretaker to grow and
thrive. If you’re reading this, you probably love XML, and you’ll
probably never stop. But if you look outside our small devoted XML
community, the fact is that XML is not pulling in a lot of newcomers.
The community is stagnant, and keeping projects like EXPath and CXAN
closed effectively starves the community until other, more thriving
communities—who view XML as a dying technology, not worth learning—
gain momentum.

What are some of these other communities? GitHub. HTML5. JQuery.
Node.js. CSS3. SASS. Stylus. The W3C might have a vicegrip on
HTML5 and CSS3, but the communities out here on the web *doing* stuff
with HTML5 and CSS3 are open, alive, and pulling in new people. And
with more people, the projects grow faster and benefit from more
diverse ideas, testing, analysis, and discourse. And communities like
GitHub are proving that a lower barrier-of-entry draws more people to
your cause and encourages a similar positive feedback-loop.

So, ask yourself:

Is the best way to get people interested in EXPath and CXAN to prevent
direct contribution?

Is the best way to convince people that EXPath and CXAN are worth
their time to present them with websites that have hardly changed in 3
years?

Is the best way to encourage growth and development of EXPath and CXAN
to guard them jealously—or to share and cooperate in building a
wonderful ecosystem of ideas, tools, and people?

Christian Grün

unread,
Apr 16, 2012, 4:46:16 PM4/16/12
to exp...@googlegroups.com
To be honest.. I agree. I believe that we all would benefit from an
open platform which can be accessed and modified by all potential
members of the community, and I would love to see EXPath and/or CXAN
opened to everyone. Otherwise, it is very likely (and it has already
happened) that developers and users will finally resign and move on to
other solutions.

Just my two cents,
Christian
___________________________

> --
> You received this message because you are subscribed to the Google Groups "EXPath" group.
> To post to this group, send email to exp...@googlegroups.com.
> To unsubscribe from this group, send email to expath+un...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/expath?hl=en.
>

Chris Maloney

unread,
Apr 16, 2012, 6:02:18 PM4/16/12
to exp...@googlegroups.com
+1

David Lee

unread,
Apr 16, 2012, 6:40:00 PM4/16/12
to exp...@googlegroups.com
-1
I've been following this project and group for a few years and admit I haven't actually used it or contributed other than by discussion, but I feel I should mention that as a leader/owner of several open source projects that in my opinion more simultaneous contributors is not necessarily a good thing. There are many metaphors to describe different kinds of open source but one which sounds hokey but is the best I've heard to describe this kind of project (and the kind I tend to run) is "It is a cathedral not a bazar".
I am not sure, but I'd be willing to bet a case of your favorite beverage that if you wanted to contribute all you'd have to do is send Florence the code and he'd look it over and if it passed muster add it. And if it didn't then maybe it shouldn't be in there.
What additional advantage is there to "open the floodgates" ? If you really have something to contribute then why not just contribute ? If your finding your contributions are being ignored or rejected then we can discuss them on this list. But until we find people are actually contributing and their effort is being ignored or rejected then its pointless to ask to open the code ...
That's my opinion ... today. Ask me tomorrow I may have changed my mind :)


----------------------------------------
David A. Lee
dl...@calldei.com
http://www.xmlsh.org


> -----Original Message-----
> From: exp...@googlegroups.com [mailto:exp...@googlegroups.com] On
> Behalf Of Chris Maloney
> Sent: Monday, April 16, 2012 6:02 PM
> To: exp...@googlegroups.com
> Subject: Re: [expath] Can we directly contribute yet?
>
> +1
>
> On Mon, Apr 16, 2012 at 4:46 PM, Christian Grün
> <christi...@gmail.com> wrote:
> > To be honest.. I agree. I believe that we all would benefit from an
> > open platform which can be accessed and modified by all potential
> > members of the community, and I would love to see EXPath and/or CXAN
> > opened to everyone. Otherwise, it is very likely (and it has already
> > happened) that developers and users will finally resign and move on to
> > other solutions.
> >
> > Just my two cents,
> > Christian
> > ___________________________
> >
> > On Mon, Apr 16, 2012 at 7:47 PM, T. Goodchild <tgood...@gmail.com>
> wrote:

> >> So.Florent is the only person with repository access.  First of all,


> >> I'd like to say that Florent is much smarter than I am, and will
> >> probably always be my technically superior to me in every way.  I
> >> think the EXpath and CXAN projects are exactly what the XML world
> >> needs, and we have Florent to thank for both of them. :)
> >>
> >> However, I have some serious concerns about the future of these
> >> projects.
> >>

> >> . The EXPath website is now over 3 years old (launched 2009-03-29).
> >>
> >> . On the EXPath "Specs" page [ http://www.expath.org/specs ], of the 8


> >> modules listed only half (4) have a release date sometime in the year
> >> of 2011.
> >>

> >> . The EXPath wiki similar shows no "last updated" times later than
> >> 2010.
> >>
> >> . The EXPath wiki still has no means to register for a username.  A
> >> **wiki**-a system invented for collaboration and cooperative
> >> editing.
> >>
> >> . The EXPath wiki also says that registration may be accomplished with


> >> an e-mail request to Florent.  In January of 2011, I did exactly
> >> that.  However, Florent said that I should instead send him an e-mail
> >> with the changes I'd like to make, so that he would make them for me.
> >> (1) It's frustrating to be told "send an e-mail to register" and then
> >> hear "Actually, I'm not giving anyone access."  (2) What is the point
> >> of a wiki at all if only one person can contribute?
> >>

> >> I mean no disrespect, but honestly-this is the most *closed* "open


> >> source" project I have ever seen. It doesn't matter how awesome the
> >> vision of EXPath (or CXAN) is if it's only allowed to grow at the pace
> >> of a single person's spare time.
> >>
> >> These projects are the seeds of something wonderful. But they were
> >> planted in 2009, and they need more than one caretaker to grow and
> >> thrive.  If you're reading this, you probably love XML, and you'll
> >> probably never stop.  But if you look outside our small devoted XML
> >> community, the fact is that XML is not pulling in a lot of newcomers.
> >> The community is stagnant, and keeping projects like EXPath and CXAN
> >> closed effectively starves the community until other, more thriving

> >> communities-who view XML as a dying technology, not worth learning-


> >> gain momentum.
> >>
> >> What are some of these other communities?  GitHub.  HTML5.  JQuery.
> >> Node.js.  CSS3.  SASS.  Stylus.  The W3C might have a vicegrip on
> >> HTML5 and CSS3, but the communities out here on the web *doing* stuff
> >> with HTML5 and CSS3 are open, alive, and pulling in new people.  And
> >> with more people, the projects grow faster and benefit from more
> >> diverse ideas, testing, analysis, and discourse.  And communities like
> >> GitHub are proving that a lower barrier-of-entry draws more people to
> >> your cause and encourages a similar positive feedback-loop.
> >>
> >> So, ask yourself:
> >>
> >> Is the best way to get people interested in EXPath and CXAN to prevent
> >> direct contribution?
> >>
> >> Is the best way to convince people that EXPath and CXAN are worth
> >> their time to present them with websites that have hardly changed in 3
> >> years?
> >>
> >> Is the best way to encourage growth and development of EXPath and CXAN

> >> to guard them jealously-or to share and cooperate in building a

daniela florescu

unread,
Apr 16, 2012, 6:48:28 PM4/16/12
to exp...@googlegroups.com
+1

On Apr 16, 2012, at 3:40 PM, David Lee wrote:

> -1
> I've been following this project and group for a few years and admit I haven't actually used it or contributed other than by discussion, but I feel I should mention that as a leader/owner of several open source projects that in my opinion more simultaneous contributors is not necessarily a good thing. There are many metaphors to describe different kinds of open source but one which sounds hokey but is the best I've heard to describe this kind of project (and the kind I tend to run) is "It is a cathedral not a bazar".

In *average*, a bazar scales better.

> I am not sure, but I'd be willing to bet a case of your favorite beverage that if you wanted to contribute all you'd have to do is send Florence the code and he'd look it over and if it passed

Passed what !? What's the condition of "passing " ?


> muster add it. And if it didn't then maybe it shouldn't be in there.
> What additional advantage is there to "open the floodgates" ?

Many. More water coming ? :-)


> If you really have something to contribute then why not just contribute ? If your finding your contributions are being ignored or rejected then we can discuss them on this list.

People did. HTTP. Zip.

A single point of failure in such an ambitious architecture is a very bad idea.


Best
Dana

David Lee

unread,
Apr 16, 2012, 6:52:35 PM4/16/12
to exp...@googlegroups.com
>
> People did. HTTP. Zip.
>
> A single point of failure in such an ambitious architecture is a very bad idea.
>
>

I'll bite. Are you saying that HTTP and zip modules were contributed but not added to the project ? Was there any response ? Has anyone suggested/volunteered adding/helping on to the management of the project short of the binary "open it to all callers" ?

daniela florescu

unread,
Apr 16, 2012, 6:57:54 PM4/16/12
to exp...@googlegroups.com
Yes, there was a lot of push back on the design of HTTP, for example.

There is still disagreement on zip.


EXpath has no good disagreement resolution process, other then "Florent didn't agree".

And after that, what do you do !?

===

No offense to anybody, especially not to Florent who spent so much time and effort on this great project,
but that's not how good, solid standards get designed.

The wisdom of crowds, while not perfect, is still the wisdom of crowds...

Best
Dana

David Lee

unread,
Apr 16, 2012, 7:10:08 PM4/16/12
to exp...@googlegroups.com
Thanks for the explanation. So it appears there is not a problem with bandwidth,
but rather a disagreement of management and decision making process.
I disagree still that opening the code inherently improves that. It might increase the quantity of 'water' or whatever fluid is going through the floodgates :) ... But that doesn't necessarily make it better - 'wisdom of crowds' or not. I will still side with the original author.

I remember a friend commenting onceaponatime when I mentioned I had an open mind ... he said "I prefer an 'unlocked mind', with an open mind who knows what people might put in there !"

----------------------------------------
David A. Lee
dl...@calldei.com
http://www.xmlsh.org


> -----Original Message-----
> From: exp...@googlegroups.com [mailto:exp...@googlegroups.com] On
> Behalf Of daniela florescu
> Sent: Monday, April 16, 2012 6:58 PM
> To: exp...@googlegroups.com
> Subject: Re: [expath] Can we directly contribute yet?
>

Adam Retter

unread,
Apr 16, 2012, 7:43:33 PM4/16/12
to exp...@googlegroups.com

Plus one from me. Florent your great and all but I think you need to make this a community, too many offers of contributions from me and others have gone unanswered!

Adam Retter

unread,
Apr 16, 2012, 7:44:58 PM4/16/12
to exp...@googlegroups.com

I have offered to help both in admin and technical contributions, but it has never gone anywhere...

Jonathan Robie

unread,
Apr 16, 2012, 8:21:22 PM4/16/12
to exp...@googlegroups.com
I would also be willing to help.

I don't think we have a defined process. We don't know how a decision is made. If two modules that do exactly the same thing are contributed, we don't know how to pick one over the other, we don't know how people get involved ...

I do think Florent has generally been doing a very good job, but as things currently stand, this isn't an open process. I suspect a lot of this is a simple bandwidth issue.

Jonathan

Michael Kay

unread,
Apr 17, 2012, 4:07:14 AM4/17/12
to exp...@googlegroups.com
I really understand Florent's difficulties with this. Opening up is
hard; it's hard to create a community of people who are able to work
together as a team, who cooperate without fighting, whose interaction
raises the quality and speeds things up rather than slowing things down.
I've never succeeded in doing this myself; for example I have accepted
very few contributions to Saxon because the quality isn't good enough or
they don't fit my vision for the strategic direction. When a community
does get established and work as a team, it's brilliant; but it's very
hard to engineer.

Michael Kay
Saxonica

On 16/04/2012 18:47, T. Goodchild wrote:
> So�Florent is the only person with repository access. First of all,
> I�d like to say that Florent is much smarter than I am, and will


> probably always be my technically superior to me in every way. I
> think the EXpath and CXAN projects are exactly what the XML world
> needs, and we have Florent to thank for both of them. :)
>
> However, I have some serious concerns about the future of these
> projects.
>

> � The EXPath website is now over 3 years old (launched 2009-03-29).
>
> � On the EXPath �Specs� page [ http://www.expath.org/specs ], of the 8


> modules listed only half (4) have a release date sometime in the year
> of 2011.
>

> � The EXPath wiki similar shows no �last updated� times later than
> 2010.
>
> � The EXPath wiki still has no means to register for a username. A
> **wiki**�a system invented for collaboration and cooperative
> editing.
>
> � The EXPath wiki also says that registration may be accomplished with


> an e-mail request to Florent. In January of 2011, I did exactly
> that. However, Florent said that I should instead send him an e-mail

> with the changes I�d like to make, so that he would make them for me.
> (1) It�s frustrating to be told �send an e-mail to register� and then
> hear �Actually, I�m not giving anyone access.� (2) What is the point


> of a wiki at all if only one person can contribute?
>

> I mean no disrespect, but honestly�this is the most *closed* �open
> source� project I have ever seen. It doesn�t matter how awesome the
> vision of EXPath (or CXAN) is if it�s only allowed to grow at the pace
> of a single person�s spare time.


>
> These projects are the seeds of something wonderful. But they were
> planted in 2009, and they need more than one caretaker to grow and

> thrive. If you�re reading this, you probably love XML, and you�ll


> probably never stop. But if you look outside our small devoted XML
> community, the fact is that XML is not pulling in a lot of newcomers.
> The community is stagnant, and keeping projects like EXPath and CXAN
> closed effectively starves the community until other, more thriving

> communities�who view XML as a dying technology, not worth learning�


> gain momentum.
>
> What are some of these other communities? GitHub. HTML5. JQuery.
> Node.js. CSS3. SASS. Stylus. The W3C might have a vicegrip on
> HTML5 and CSS3, but the communities out here on the web *doing* stuff
> with HTML5 and CSS3 are open, alive, and pulling in new people. And
> with more people, the projects grow faster and benefit from more
> diverse ideas, testing, analysis, and discourse. And communities like
> GitHub are proving that a lower barrier-of-entry draws more people to
> your cause and encourages a similar positive feedback-loop.
>
> So, ask yourself:
>
> Is the best way to get people interested in EXPath and CXAN to prevent
> direct contribution?
>
> Is the best way to convince people that EXPath and CXAN are worth
> their time to present them with websites that have hardly changed in 3
> years?
>
> Is the best way to encourage growth and development of EXPath and CXAN

> to guard them jealously�or to share and cooperate in building a

Alain Couthures

unread,
Apr 17, 2012, 4:48:58 AM4/17/12
to exp...@googlegroups.com, Michael Kay
+1

The same for XSLTForms!

Alain Couthures
agenceXML

Le 17/04/2012 10:07, Michael Kay a �crit :

David Carlisle

unread,
Apr 17, 2012, 7:14:50 AM4/17/12
to exp...@googlegroups.com
The original message mentioned both expath and cxan,

expath might be considered a specific coding project so as people have
said, there can reasonably be differing views as to how "open" it is,
but for cxan if it is to be at all comprehensive (and live up to
stated analogy with cpan or ctan) then it has to be essentially wide
open. CTAN (the first of these Comprehensive ??? Archive Networks)
just really requires that there is a clear statement of the licence
under which a package is to be distributed and then it is added. The
archive maintainers maintain a light hand encouraging packages to be
submitted to reasonable locations within the archive and preferably
with some documentation but the actual rules are pretty simple, the
instructions for uploading to CTAN are here

http://www.tex.ac.uk/upload/

David

Dave Pawson

unread,
Apr 17, 2012, 7:20:07 AM4/17/12
to exp...@googlegroups.com
On 17 April 2012 12:14, David Carlisle <d.p.ca...@gmail.com> wrote:
> The original message mentioned both expath and cxan,
>
> expath might be considered a specific coding project so as people have
> said, there can reasonably be differing views as to how "open" it is,
> but for cxan if it is to be at all comprehensive (and live up to
> stated analogy with cpan or ctan) then it has to be essentially wide
> open.

I can see the advantages of that model David. Couple of points.
1. This fails when spammers take a hand? How does cpan handle that?
2. If I have n zip packages... how to select/reject one?
3. I thought that expath was following the ideas of exslt, which was
to seek user input, for addition to XSLT/XPATH n+1, which would
favour a single package definition? Again, n packages goes against
this idea to some extent.
Is CPAN a special purpose host, written for CPAN?

I'm with Mike Kay on this, a bazaar model seems an uncomfortable fit somehow.

regards

--
Dave Pawson
XSLT XSL-FO FAQ.
Docbook FAQ.
http://www.dpawson.co.uk

David Lee

unread,
Apr 17, 2012, 7:21:43 AM4/17/12
to exp...@googlegroups.com
I was thinking about that last night.
If CXAN is to be like CPAN it does need to be reasonably open.
But I think there is a difference, philosophy. One that should be clarified.
Is CXAN supposed to be like CPAN or is it supposed to be 'Authoritative'.
As an example ... suppose I write a library for processing CSV.
Is there supposed to be one and only one CSV library in CXAN ? If so then there needs to be an arbitrator which decides "Which is the CSV Library I want to bless". This will likely be restrictive and subject to the goals and philosophy of the arbitrator.

But if there can be any number of them then it should be more open.
From my memory of reading this newsgroup I get the feeling that CXAN is more like the former. Perhaps to the misunderstanding of many (myself included).


----------------------------------------
David A. Lee
dl...@calldei.com
http://www.xmlsh.org

> -----Original Message-----
> From: exp...@googlegroups.com [mailto:exp...@googlegroups.com] On
> Behalf Of David Carlisle
> Sent: Tuesday, April 17, 2012 7:15 AM
> To: exp...@googlegroups.com
> Subject: Re: [expath] Can we directly contribute yet?
>

David Carlisle

unread,
Apr 17, 2012, 7:38:34 AM4/17/12
to exp...@googlegroups.com
On 17 April 2012 12:20, Dave Pawson <dave....@gmail.com> wrote:
> On 17 April 2012 12:14, David Carlisle <d.p.ca...@gmail.com> wrote:
>> The original message mentioned both expath and cxan,
>>
>> expath might be considered a specific coding project so as people have
>> said, there can reasonably be differing views as to how "open" it is,
>> but for cxan if it is to be at all comprehensive (and live up to
>> stated analogy with cpan or ctan) then it has to be essentially wide
>> open.
>
> I can see the advantages of that model David. Couple of points.
> 1. This fails when spammers take a hand? How does cpan handle that?

CPAN (CTAN's younger brother) I can't speak to but CTAN is explicitly
_not_ designed to be the authoritative reference collection. It is
supposed to be _C_omprehensive. Names are allocated on a
first-come-first-served basis, and as long as the name of your package
is different the fact that there is a package offering similar
functionality is not a problem, in fact it is encouraged, people often
post enhanced or re-implemented versions of things.

Spammers may (and probably do) bother the archive maintainers but they
have methods to handle it (Robin's be doing it for over 20 years:-)
Spam doesn't get on to the site, upload isn't automatic there is
always a human in the loop.

> 2. If I have n zip packages... how to select/reject one?

You don't.

> 3. I thought that expath was following the ideas of exslt, which was
> to seek user input, for addition to XSLT/XPATH n+1, which would
> favour a single package definition? Again, n packages goes against
> this idea to some extent.

My comments were specifically about CXAN. EXpath has different considerations.

> Is CPAN a special purpose host, written for CPAN?

I am not sure I understand that question.


>
> I'm with Mike Kay on this, a bazaar model seems an uncomfortable fit somehow.

cathedral v bazaar is talking about software development. C?AN is not
about software development it is an _Archive_ and supposed to be
_Comprehensive_ (which accounts for half the letters in its name) in
otherwords it should just have everything (as far as practicable)
>
> regards
>

David
(I was somewhat involved in the original discussions that led to
setting up CTAN, but never actually helped with the maintenance of the
archive)

Dave Pawson

unread,
Apr 17, 2012, 7:53:16 AM4/17/12
to exp...@googlegroups.com
On 17 April 2012 12:38, David Carlisle <d.p.ca...@gmail.com> wrote:

>> 2. If I have n zip packages... how to select/reject one?
> You don't.

So for purpose X, there may be n packages, I'd have thought expath 'users'
might like some sort of way of choosing?
Is that not an issue in CTAN? you only see n different packages
which may/may not be named coherently. -1 from a user view
just seeking a package to do, say , zip?

>
>> 3. I thought that expath was following the ideas of exslt, which was
>> to seek user input, for addition to XSLT/XPATH n+1, which would
>> favour a single package definition? Again, n packages goes against
>> this idea to some extent.
>
> My comments were specifically about CXAN.  EXpath has different considerations.

Noted. I think they are different in a number of aspects.

>
>> Is CPAN a special purpose host, written for CPAN?
>
> I am not sure I understand that question.

From your comments re spam I thought perhaps there
was some way of rejecting software. It seems Robin
has that job.
Is Florent up for that level of filtering?


>>
>> I'm with Mike Kay on this, a bazaar model seems an uncomfortable fit somehow.
>
> cathedral v bazaar is talking about software development. C?AN is not
> about software development it is an _Archive_ and supposed to be
> _Comprehensive_ (which accounts for half the letters in its name) in
> otherwords it should just have everything (as far as practicable)

I think something akin to that is relevant. The comments were
that 'Florent rules' I read as somewhat negative, despite his
decisions getting support.

Florent and Robin clearly have different jobs!

I guess the XSLT WG would have a job on saying yes, we want to include
a 'package' to do (say zip), and then have to choose from n different
implementations? Just as users? Over time that is likely to either resolve
itself (popularity perhaps) or produce a fork aspect to expath?

David Carlisle

unread,
Apr 17, 2012, 8:33:51 AM4/17/12
to exp...@googlegroups.com
On 17 April 2012 12:53, Dave Pawson <dave....@gmail.com> wrote:
> On 17 April 2012 12:38, David Carlisle <d.p.ca...@gmail.com> wrote:
>
>>> 2. If I have n zip packages... how to select/reject one?
>> You don't.
>
> So for purpose X, there may be n packages, I'd have thought expath 'users'
> might like some sort of way of choosing?
> Is that not an issue in CTAN?

No. It is an issue for TeX distributions (such as texlive or miktex
etc as they have to decide what to install by default and what to
offer at all as part of the distribution. It is not an issue for the
archive which aims to be comprehensive so _by design_ includes the
good the bad and the ugly.

> you only see n different packages
> which may/may not be named coherently. -1 from a user view
> just seeking a package to do, say , zip?

users typically use a coherent distribution like texlive where someone
has made that choice. If they don't like the choice they can browse
around CTAN and see essentially every TeX package that an author has
thought worth archiving in a public place for the last 25 years.

>
>>
>>> 3. I thought that expath was following the ideas of exslt, which was
>>> to seek user input, for addition to XSLT/XPATH n+1, which would
>>> favour a single package definition? Again, n packages goes against
>>> this idea to some extent.
>>
>> My comments were specifically about CXAN.  EXpath has different considerations.
>
> Noted. I think they are different in a number of aspects.

Saying that they are different in a number of aspects misses the main
point that they are not comparable at all.

>
>
>
>>
>>> Is CPAN a special purpose host, written for CPAN?
>>
>> I am not sure I understand that question.
>
> From your comments re spam I thought perhaps there
> was some way of rejecting software. It seems Robin
> has that job.

No software is never rejected (on software grounds) The software is
not run or tested or necessarily even looked at. If the submission has
a ctan compatible licence and claims to be relevant to tex then it
will go on the archive. That's what the C in the name means.

>>>
>>> I'm with Mike Kay on this, a bazaar model seems an uncomfortable fit somehow.
>>
>> cathedral v bazaar is talking about software development. C?AN is not

>


> I guess the XSLT WG would have a job on saying yes, we want to include
> a 'package' to do (say zip), and then have to choose from n different
> implementations? Just as users? Over time that is likely to either resolve
> itself (popularity perhaps) or produce a fork aspect to expath?
>

that's about expath rather than cxan , but as I said in my original
reply my comments were only about cxan,

David

Florent Georges

unread,
Apr 18, 2012, 7:49:12 PM4/18/12
to exp...@googlegroups.com
On 16 April 2012 19:47, T. Goodchild wrote:

Hi,

> Is the best way to get people interested in EXPath and CXAN to
> prevent direct contribution?

Thank you all for your comments on this subject! First of all, I'd
like to say this is not the intent, of course, and to draw as well a
precise line between different projects and different facets of EXPath
mentioned here:

- EXPath itself is a collection of specifications

- the EXPath website contains a wiki

- CXAN is a separate website/system, even though it is based on
the packaging system defined by EXPath

- there are several implementations of several EXPath specs, some
open source, some not, some from me, some from other people/
organizations/companies

I won't cover the latter here. It is not out of the scope of this
mailing list, but I think it is out of the scope of this specific
thread.

Before detailing how I see collaboration for the specs, the wiki,
and CXAN, I'd like to add that I'd really love to see more people
collaborate actively, and I am sure I could improve some processes
and/or tools. If you have specific ideas, please share them!
(probably in a new thread) By specific, I mean something else than
the usual "I have a wonderful idea, I want it in EXPath, but I am not
prepared to discuss it before, I just want to add it directly and to
have full power even before showing a few times that I understood the
project and how to collaborate with the community"...

** Specifications

The specifications are the core of EXPath. They ARE EXPath. The
way I designed the process for writing and releasing them is quite
vague, on purpose. In order for it to be more agile. Each spec has
an editor (sometimes a few, but only one is the editor, others are
referred to as contributors).

The idea is that the editor is the one guy responsible for the spec.
Responsible to lead the conversations, to respond to comments and
questions, and to integrate feedback into the spec, with the goal to
have a 1.0 version of the spec. I am the editor for some specs,
others are the editors of other specs. In the latter case, I take
part to the discussions as any other one, but I let the editors the
final decision, even when I don't agree with a design choice. I want
the editors to be the one who decide, in order to be consistent in
each single one spec.

Some editors proved to be quite autonomous and efficient, deeply
involved in the community and discussions. Others were less. I don't
blame anyone of course, and I want to thank the whole community and
the editors especially for all their important work!

Off course, before publishing the first draft of a new module, it
must comply to some requirements. From technicals to editorials. But
I try to guide them as much as possible. Maintaining your own spec
require some investment in time, but it is really worth it if you
believe in (even loose) standardization!

I would really like to see more people offering to be editors for
new specs they'd like to have in EXPath. In the past, some people had
good ideas of new modules, but were not prepared to go through the
editorial process themselves and at the same time were frustrated no
one would do it for them. I am not sure this is the right approach...

** Wiki

I understand the frustration of some people regarding the fact the
wiki is not completely open. This is because I saw so much cases
where the wiki quickly evolved into something where no one could find
it way, without any consistency, and at the end of the day not used by
anyone.

It does not seems to me that it's not reasonable to ask people to go
through emails the first couple of times, before being granted direct
write access once they have proved themselves to understand the
editorial guidelines...

But it seems to have been a psychological obstacle as only a few
people offered contributions. If anyone wants to take over the
responsibility of the wiki, I think this is maybe a good opportunity
to do so. I can host it on the EXPath website, but I think there are
enough free solutions available out there for someone to start a new
wiki, open it (because it seems this is the one issue with the
existing wiki), maintain it, and manage it. If this is successful,
I will link it (or integrate it somehow) from the EXPath website
instead of the current wiki.

Those who want an open wiki, it's time to take over your own
destiny ;-)

** CXAN

CXAN is not EXPath. It uses EXPath, at least its packaging system.
But I think this is a very important piece in the XML landscape.
Technically, the website still needs some improvements in order to be
ready for real collaboration.

But regarding the very, very few requests to add a new package to
CXAN, I really don't think this is an issue, yet.

One of the central and important feature of CXAN is that it gives a
unique short name to each package, where technically the packaging
system is based on URIs. URIs are good to "guarantee" unique names
even though there is no authoritative source assigning them, and they
are good for programs. Short names are better for humans, but they
need an authority to manage and assign them.

So the idea is that the maintainer of a package (its implementer for
instance) must ask to add the package to CXAN. This is not at all in
order to check the content or the quality of the package (but maybe
the packaging itself), just to validate the short name (the CXAN ID)
for that package. Because that ID is aimed to be persistent. Once
the ID has been assigned, the idea is that the maintainer can edit and
update new version of that package. For the better and the worse.

The system itself is also design from the beginning to be able to
setup another repository, so instead of or in addition to the
"official" CXAN repository at cxan.org, one can setup another repo
(e.g. local to a company or to address a specific domain or to build a
specific "distribution"...)

If you have a library/application/set of schemas/anything else that
you'd like to add to CXAN, please tell me. I can help you (and
actually already did) to build a package out of it if you don't
already did and are not used to the packaging system. Even if you are
not the author of the library, you can be the package maintainer on
CXAN (providing you are not violating any licensing of course...) if
you want to benefit from having it as a package and on CXAN (for
instance some de-facto standard schemas or queries or pipelines or
stylesheets).

** Conclusion

I want to base the collaboration in EXPath on responsibility. Over
single units. A single spec in EXPath. A single page or section in
the wiki (or the full wiki if someone is ready to do so). A single
package in CXAN.

If you want to take responsibility over a new spec, over the wiki or
to add a new package in CXAN, please tell me so in private or in this
mailing list! And if you have any specific suggestion to improve
collaboration, please tell us so! (but I won't open everything to
everyone, especially the specs, because they need to be rather stable)

If you want to continue this discussion, it is maybe a good idea to
start a new thread regarding the nature of your comments (about specs,
CXAN, the wiki, the EXPath processes, etc.)

Thank you all for this discussion and for your ideas!

--
Florent Georges
http://fgeorges.org/
http://h2oconsulting.be/

Max Toro

unread,
Apr 18, 2012, 8:37:56 PM4/18/12
to exp...@googlegroups.com
>  The idea is that the editor is the one guy responsible for the spec.
> Responsible to lead the conversations, to respond to comments and
> questions, and to integrate feedback into the spec, with the goal to
> have a 1.0 version of the spec.  I am the editor for some specs,
> others are the editors of other specs.  In the latter case, I take
> part to the discussions as any other one, but I let the editors the
> final decision, even when I don't agree with a design choice.  I want
> the editors to be the one who decide, in order to be consistent in
> each single one spec.

Having one guy responsible for a spec is the reason we have 3 year old
drafts. Guy gets busy, spec freezes, everyone looses interest.

--
Max Toro

Christian Grün

unread,
Apr 19, 2012, 6:41:42 AM4/19/12
to exp...@googlegroups.com
Florent,

thanks for your feedback. The current thread shows very well that we
all would like to see EXPath flourish. My personal impression is that
the progress has rather been stagnant in the last years, which is why
we have stopped implementing the remaining EXPath modules, and
defining our own in collaboration with others.

As both EXPath and CXAN are your very own projects, all I want to do
is write about my own experiences. Since the beginning of BaseX, we've
been asking for external contributions, and the most promising way to
let this happen was to provide more open platforms. Our documentation
is now placed in a Wiki; everyone who's registering gets an account,
and the registration process has only been introduced to stop obvious,
semi-automatical, spamming activities. We observed that all of the
external contributions have been helpful to improve our documentation,
as most users still show great respect before editing public contents.
Next, our source code is now placed on GitHub; many developers are
forking the code, and we are getting more and more patch proposals. Of
course, this platform is not completely free, as we can always choose
which patches (aka pull requests) to adopt in our main branch (and
this makes sense; here I agree with Michael Kay), but the contribution
barrier is much lower if you can act independently (here I am in line
with Chris Maloney).

My belief is that open platforms get more and more mandatory nowadays
to attract high-prolific contributors. There is no doubt that this
will potentially open doors to flawed submissions, too, but those
contributions will eventually harm the contributors, and not the
platform itself. After all, this is already the case: better EXPath
specifications have been implemented by more projects than others.

Currently, I can't imagine that more open platforms for XQuery will
lead to an explosive growth of contributions; the community seems too
small for that. But I may be wrong, and things may change over time
(to be honest, I still can't grasp that Wikipedia has managed to get
so many serious users to outstrip all other commercial dictionaries --
and Jimmy Wales as its BDFL is much more well-known than ever before).

Christian

PS: Regarding your appeal to get in contact with you, Florent, I'm
sorry to say that all my recent mails to you have gone unanswered.
I've got similar reports from others. This is by no means meant to
blame you; I just assume you are extremely busy.

Tony R.

unread,
Apr 19, 2012, 12:14:09 PM4/19/12
to exp...@googlegroups.com
On Apr 18, 2012, at 8:37 PM, Max Toro wrote:
> Having one guy responsible for a spec is the reason we have 3 year old
> drafts. Guy gets busy, spec freezes, everyone looses interest.
>
> --
> Max Toro


Exactly! Well said. :)


—Tony


Tony R.

unread,
Apr 19, 2012, 12:14:23 PM4/19/12
to exp...@googlegroups.com
On Apr 19, 2012, at 6:41 AM, Christian Grün wrote:
Next, our source code is now placed on GitHub; many developers are
forking the code, and we are getting more and more patch proposals. Of
course, this platform is not completely free, as we can always choose
which patches (aka pull requests) to adopt in our main branch (and
this makes sense; here I agree with Michael Kay), but the contribution
barrier is much lower if you can act independently (here I am in line
with Chris Maloney).


THIS.  +100 times this.  

When I started this thread, this was the ideal situation I had in mind.  Several people raised concerns about the “free-for-all” nature of some projects, and how it can lead to a worse signal-to-noise ratio for the projects.  I never intended to suggest that these projects would benefit from being “open” as in “anarchy”.  Structure is good.  Benevolent dictators are (IMO) usually a good thing.  

Don’t believe me?  See Jimmy Wales (thanks Christian).  Before WikiPedia, Jimmy Wales created something called NuPedia.  It was similar, except that it had strong limitations on which articles it accepted.  

During its 2-year lifetime, NuPedia only produced 24 articles.  That’s an average of 1 article per month.  

And what was NuPedia’s rationale for such tight restrictions?  Pretty much exactly what Florent said are the reasons he has kept these projects closed.

Having decided that NuPedia wasn’t working, Jimmy Wales decided that a lower barrier of entry was the way to go.  So he created WikiPedia—and the rest is Internet History.  


Currently, I can't imagine that more open platforms for XQuery will
lead to an explosive growth of contributions; the community seems too
small for that.

Exactly.  

With such a small community, we’re already at a disadvantage when it comes to collaborating.  A smaller community means there are fewer people to spread the work around, fewer people to contribute to a discourse with diverse ideas, fewer people to test for bugs—and fewer people overall, so that statistically our community is less likely to contain a significant number of “superstar” developers.  

Keeping one person as the single bottleneck for all dialogue makes that limitation even worse.  


PS:  Regarding your appeal to get in contact with you, Florent, I'm
sorry to say that all my recent mails to you have gone unanswered.
I've got similar reports from others.


Yup.  Ditto.  


This is by no means meant to
blame you; I just assume you are extremely busy.


Also ditto.  Florent, nobody expects you to be able to answer everything all by yourself.  So, for starters: why not choose a few people you trust to oversee different parts of the projects?  


—Tony

mozer

unread,
Apr 19, 2012, 12:57:28 PM4/19/12
to exp...@googlegroups.com

mozer

unread,
Apr 20, 2012, 7:16:01 AM4/20/12
to exp...@googlegroups.com

Rushforth, Peter

unread,
Apr 23, 2012, 10:40:12 AM4/23/12
to exp...@googlegroups.com
Hi,

I am on a REST kick, so I have been thinking about URIs a lot lately :-).

There is a new RFC published, which has pretty solid credentials ;->, for URI Templates
http://tools.ietf.org/html/rfc6570

It occurred to me that this would be a useful target for a set of functions in xpath,
where the use case is that the xpath-enabled client receives an XML document that
has URI templates in its content, as well as values embedded in the xml somewhere.

The objective would be to generate legal URIs given the templates (as strings) and the input values,
possibly as sequences.

Unfortunately, I did not notice when the last call for xpath 3 requirements came and went.
However, since the RFC was only recently published, seems like it might have been
premature to base requirements on a draft.

Now that it's published, maybe we could discuss a function library for expath here, with
possible eventual inclusion in xpath proper.

Any thoughts? Would this be of interest or value to anyone?

Cheers,
Peter


Adam Retter

unread,
Apr 23, 2012, 11:43:23 AM4/23/12
to exp...@googlegroups.com
I think such functions could already be written in pure XQuery or
XSLT, so an EXPath function library is not really nessecary.

On 23 April 2012 15:40, Rushforth, Peter

> --
> You received this message because you are subscribed to the Google Groups "EXPath" group.
> To post to this group, send email to exp...@googlegroups.com.
> To unsubscribe from this group, send email to expath+un...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/expath?hl=en.
>

--
Adam Retter

skype: adam.retter
tweet: adamretter
http://www.adamretter.org.uk

mozer

unread,
Apr 24, 2012, 5:27:15 AM4/24/12
to exp...@googlegroups.com
Let's say it otherwise

If you want to be able to better contribute, you may want to join the W3C community group


Xmlizer

Jim Fuller

unread,
Apr 19, 2012, 6:53:57 AM4/19/12
to exp...@googlegroups.com
I've recently launched a little tool for managing deps in the XML world … its somewhat tangentially related to this conversation thread and expath packaging.

What is depx ?

      Its the simplest package manager for XML tech I needed, it does package management via 'convention' versus invoking 'black magic' and no doubt there are gaps in its functionality

Whats depx main goal

      Its mainly an experiment to inspire packaging efforts in the XML domain as well as show people how much great software there exists for xml

What can go in the depx repository ?

      xquery
      xslt
      xproc
      schema
      xml
      js
      css

At this moment its best for library orientated stuff, rather then full blown apps (but maybe in the future … ).

Check out the website, which enumerates the various packages

      http://depx.org/

I am helping people add to the repo (or if you are github conversant, its just a matter of forking depx and pull request).

More information can be found at the github site which also acts as the depx package repository

      https://github.com/xquery/depx

Things to do & test

      *  browse http://depx.org and try to install a package ( hint: you will need to install depx client )

      * if you have a package already in depx, fork depx, update your package and send a pull request

      * add a new package (via forked depx) and send pull request

Please send comments, criticisms and suggestions to the issue tracker for depx here

    https://github.com/xquery/depx/issues

as for its relation to expath ... I think Florent's heroic efforts should be commended ... and I see the work I am doing with depx supplants expath as I plan to offer expath xar's at some point.

Jim Fuller

Adam Retter

unread,
Sep 4, 2012, 1:19:06 PM9/4/12
to exp...@googlegroups.com, christi...@gmail.com, Florent Georges
Was there ever any resolution from this?

As I am about to write a number of function modules which would
ideally come under the 'EXPath' banner, I wonder how I can do this?
There is no mechanism for me to fork EXPath specs, create my spec, and
issue a pull request.

Florent,

Any chance you could move all the EXPath stuff to GitHub and set it up
so people can collaborate. I think Editors of a spec, for which there
should be a repo for each I guess, should be able to accept pull
requests, whereas contributors just fork and send pull requests.
In fact if you like, I can do it for you as I just did it for the
EXQuery project and RESTXQ spec?

Let me know...
> --
> You received this message because you are subscribed to the Google Groups "EXPath" group.
> To post to this group, send email to exp...@googlegroups.com.
> To unsubscribe from this group, send email to expath+un...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/expath?hl=en.
>



Florent Georges

unread,
Sep 5, 2012, 9:41:06 AM9/5/12
to exp...@googlegroups.com
On 4 September 2012 19:19, Adam Retter wrote:
> ??? wrote:

Hi,

>> Any chance you could move all the EXPath stuff to GitHub and set it
>> up so people can collaborate. I think Editors of a spec, for which
>> there should be a repo for each I guess, should be able to accept
>> pull requests, whereas contributors just fork and send pull
>> requests.

> Was there ever any resolution from this?

Yes, see https://github.com/fgeorges/expath-cg, which is the
official source repository for the W3C EXPath Community Group (see
http://w3.org/community/expath/). All the spec-related activity
should now go through the CG [*], and its mailing list.

Looking forward to your contributions!

Regards,
[*] CG stands for Community Group, not for Christian Grün ;-)

Christian Grün

unread,
Sep 5, 2012, 11:21:15 AM9/5/12
to exp...@googlegroups.com
Hi Florent,

> [*] CG stands for Community Group, not for Christian Grün ;-)

;) By the way, have you already thought about adding the other,
remaining specifications in the repository? And would you be
interested in converting EXPath into an "organization", similar to the
W3 page [1]?

Best,
Christian

[1] https://github.com/w3c
___________________________

Adam Retter

unread,
Sep 5, 2012, 12:15:50 PM9/5/12
to exp...@googlegroups.com
> ;) By the way, have you already thought about adding the other,
> remaining specifications in the repository? And would you be
> interested in converting EXPath into an "organization", similar to the
> W3 page [1]?

Yes it would be great is there was an EXPath organisation on GitHub
with its own repos, rather than just a repo under a user...

Joe Wicentowski

unread,
Sep 5, 2012, 12:21:40 PM9/5/12
to exp...@googlegroups.com
+1

Tony R.

unread,
Sep 5, 2012, 3:52:24 PM9/5/12
to exp...@googlegroups.com
On Sep 5, 2012, at 12:15 PM, Adam Retter <adam....@googlemail.com> wrote:
;) By the way, have you already thought about adding the other,
remaining specifications in the repository? And would you be
interested in converting EXPath into an "organization", similar to the
W3 page [1]?

Yes it would be great is there was an EXPath organisation on GitHub
with its own repos, rather than just a repo under a user...


+1 for me too! :)
Reply all
Reply to author
Forward
0 new messages