cocoon-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Stefano Mazzocchi <stef...@apache.org>
Subject Re: Proposal for Lenya
Date Fri, 21 Feb 2003 14:32:40 GMT
Steven Noels wrote:

> My main concern is the fact that Lenya does not come only with a 
> community, but also with a code base. That code base is in use already 
> at a selected number of commercial installations (which is good, of 
> course). I hope to be proven wrong, yet I fear the existing codebase is 
> intimately linked with these installations - hence the number of 
> publications in CVS. The main group of committers _might_ also depend 
> (financially) on supporting these installations. OK, I'm FUDing here, 
> but I have seen how some other candidate contributions were received 
> lately, and I want to make sure we don't just do the 'huuray - a CMS 
> project at Apache' thing.

Excuse me, but what apache project didn't come with a codebase?

In case you don't know, Apache *requires* an existing codebase to start, 
  there were attempts in the past to create an open source project out 
of design discussions and no code. It *NEVER* worked.

Here, we are lucky enough to have a CMS that is being used in one of the 
most important newspaper of a very high-tech country (switzerland) and 
it's based on apache technologies. Maybe it won't be perfect, but it's a 
seed for a community around CMS issues.

I really can't see what can be a better way for us to start into this.

> Please understand me: if you and some Wyona guys and some other 
> CMS-oriented folks would jump up and invite people to join a CMS 
> project, I would have a vested interest in joining that community. Now 
> if joining that community means supporting an existing, massive codebase 
> instead of proper [RT]-ing with sufficient blank space to be filled in, 
> I would feel less challenged. 

The software is what its development community wants it to be.

You know how I run communities and what are the things I value. Michael, 
who is the technical leader, resonates with my vision a lot and has been 
playing very nicely with Cocoon.

You might not know this, but years ago, Michael approached me asking if 
it was possible to use binary pipelines in cocoon and I said no and will 
probably never happened. This triggered them to write their own pipeline 
engine. Later, they found out that cocoon would have been a better 
choice, but not technically, but community-wise.

I feel that between a strong codebase and a strong community, Micheal 
will choose a strong community. This means delegation, RT-style 
refactoring and so on. The way Cocoon works. The way you like. The way I 
like.

Of course, doing so there will be users to care for and migration paths 
to develop when refactoring takes place. But this happens in Cocoon and 
doesn't seem to bother you.

> I'm not saying this will happen, and I 
> find it great that Wyona people went the extra mile and open-sourced 
> their stuff, but I'm not sure whether your grand vision of an ASF CMF 
> will escape from this (codebase) donation.

Well, I can't be sure either. This is why we are going thru incubation: 
it should be a period to estimate *IF* this codebase with these people 
is the right seed for a potential healthy community.

If now, well, at least we tried.

> Of course, I should shut up and get involved (I'll do that). Still, the 
> idea behind Incubation Proposals is that people can react upon them. I'm 
> only voicing my concerns, and I would love to be proven wrong.

I appreciate you voicing your concerns because they allow me to express 
my feelings about this.

the CMS world is a very difficult one. yet, content needs to be edited, 
managed and published. and those miriads of half-based solutions hurt me 
for lack of reuse between their components and solutions.

Leyna will hopefully bring more choices and less pain for content 
management solution builders, not the opposite. Or I will consider the 
project failed and I will vote -1 to make it exit incubation stage.

>>> * A sh*tload of (even Cocoon-based) (half-baked) CMS solutions exist 
>>> already, which might or might not be willing to join ASF in the 
>>> future. What will happen if Lenya (nice name BTW) comes and claims 
>>> that area? 
>>
>>
>>
>> "claims that area"? how can it possibly happen? Apache is not 
>> something that you go and homestead. Here we are talking about 
>> incubating a community around problems that affict many people (and 
>> you as well).
>>
>> If you have a better proposal or codebase to start from, I'm happy to 
>> hear about it.
> 
> 
> Gee - I knew that one was coming ;-)

You can't stop an effort with FUD overhere, expecially an incubation 
one. It's not fair!

> I have a customer asking me about a Cocoon-based CMS [it's a long-term 
> project, so I don't need it right now, so please don't send me quotes 
> for all your homebrewn CMS solutions, beloved colleagues ;-)]. I have a 
> number of options:
> 
>  * try to find out if Lenya fits my requirements and use it
> 
>    - be happy with it as-is (just be a user)
>    - find holes which I can help to plug
>    - wreak havoc while trying to apply Lenya to a use-case it hasn't 
> been designed for
> 
>  * pretend it doesn't exist and write my own one
> 
>    - closed source
>    - open source
>      - cocoondev.org
>      - donation to ASF
> 
> So eventually, given some time, I might have a codebase, but it is still 
> very uncertain.

Leyna is here *NOW*. This is a big plus, in my book. I really can't see 
why shouldn't be one in yours as well.

> Now at this time however, there _are_ already people with CMS codebases. 
> These people might not be willing to work on Lenya, since they are bound 
> to have a different vision and/or implementation, and working on Lenya 
> might mean killing their own baby. I'm afraid (and I repeat: please show 
> me where I'm wrong) that CMS is an area where it is very hard to come up 
> with a decent, generic framework if you don't put many people together 
> right from the start. Even then, there's only a slim chance.

Michael is one of the promoter of the Open Source Content Management 
Conference (http://www.oscom.org/). I think that very few people I've 
met have so much 'open mind' about different open source CMS solutions 
than him.

This said, there will be ego problems. I know that. It's not so hard to 
forecast. But again, that's another reason to move into incubation stage.

>>> Will it be the reference ASF CMS tool?
>>
>>
>>
>> There is no *reference* anything inside apache. It's software 
>> darwinism, the community success decides what lives and what not. 
>> Tomcat might be the 'official' servlet engine, yet Cocoon ships Jetty 
>> and everybody is happier than ever.
>>
>> The system works, it's adaptive, impartial and meritocratic.
> 
> 
> Sure, and I'm wondering out loud whether the move to Apache will be good 
> for Lenya. 

This is not what we are discussing here. We are discussing if incubating 
Lenya under Apache is good for Apache. The Wyona development community 
(with our advice) already thought about it and decided it was worth trying.

Of course, they are trying to use apache, but it's the usual "do ut des" 
(latin for 'give and take'): they earn visibility and apache earns a new 
community that can improve its overall visibility.

Since it's not certain that it will work, we are proposing the 
incubation facility exactly to avoid hurting other projects by 
bootstrapping the lenya community.

> They have a community, there's buzz, a diverse set of 
> committers so-they-say, and soon they might find themselves in the midst 
> of a much larger community. Does Wyona/Lenya needs the ASF brand to 
> succeed?

Wyona needs to find a place where there is no automatic correlation to a 
commercial entity, otherwise communities don't grow even if seeded 
because nobody likes the feeling of working for a company for free.

They could go to SF, but the Apache Incubator was created as a way to 
bring new potential seeds for communities and help them grow around the 
infamous 'apache way'.

> To attract new committers? Given OSCOM and quite some vocal 
> people working on it, I'm pretty sure Lenya receives a fair bit of 
> attention. 

Yes, but no matter how much you water your seed, it can't grow on 
concrete. You need a more fertile substrate and this is what this 
incubator is meant to provide.

And if then fails, we might learn stuff that might help future efforts 
not to fail.

> But as you say, the system is there and it works. We are 
> allowed to try and do some future-telling, aren't we?

Sure.

>>> Can CMS be considered an area where the ASF wants to operate in, like 
>>> Zope (CMF) is...?
>>
>>
>>
>> The ASF is for the evolution of web technologies. Would you state that 
>> making structured content production and management easier is a goal 
>> the ASF shouldn't deal with?
> 
> 
> Aren't we doing that yet? We have XML parsers, XSLT engines, WebDAV 
> stacks, servlet containers, and much more (like mod_perl and AxKit, Matt 
> ;-). Assembling those into a product will not necessarily evolve web 
> technology IMHO.

Well, Forrest is just a bunch of stylesheets, schemas and sitemaps on 
top of Cocoon. Are you proposing we drop Forrest because it's just 
assembling things into a product that doesn't necessarely evolve web 
technology :-)

Seriously: building a CMS on top of Cocoon is *HARD*. I've tried and the 
complexity is much bigger than it seems. I consider Wyona itself just a 
starting point, a seed, in fact.

If you start entering stuff like workflowed revisioning of content, 
multi-user annotation, repository abstraction and stuff like that, if 
clearly understand that it's not just a matter of assembling existing 
stuff, but it's *MUCH* more complex than this.

What I want is not a CMS product, but a content management framework 
backed up by a community of people that know problems and solutions and 
that will write the glue between the pieces that exist and foster 
development of those pieces that are missing.

But in order to do this, we have to start somewhere and Lenya and the 
Wyona people are the best option we could find on outthere!

[again, if you have other alternatives, I'd love to hear them, but they 
have to be here now, not on paper on a possible distant future]

>>> Or do we stick to supporting technology like servlet containers, http 
>>> stacks, build tools ... I know there is no policy at ASF that states 
>>> only one CMS project can exist under the ASF umbrella, but still 
>>> there is only one JetSpeed, one Tomcat, one Cocoon, etc etc - I hope 
>>> my point is clear.
>>
>>
>>
>> It's clear that you don't understand what we are trying to do.
> 
> 
> I'm _trying_ to understand. Which is better than silently ignoring ;-)

True and appreciated.

>> We don't want to *homestead* the CMS niche of the ASF (whatever that 
>> means in OSS terms!), we want to build a community that will focus on 
>> serious content management and, so far, there is no such community in 
>> the ASF since Slide is repository-oriented (no publishing nor editing 
>> layer) and Forrest is publishing-oriented (no repository nor editing 
>> layer).
> 
> 
> Sure thing. Since Lenya does publishing too, what is the future of 
> Forrest in your scenario? This is not a pun nor a cynical remark: I'm 
> honestly asking. It seems like you guys have spend some serious thoughts 
> in preparing this proposal, and I would like to understand the rationale 
> behind it.

I've spent at least two years thinking about the perfect content 
management framework because content management is what I'm really missing.

my work on cocoon, the spin-off of forrest, my involvement in JSR 170, 
my hopes on xindice, on xopus and now on lenya reflect that.

forrest was created as a stylebook replacer but has grown into something 
that is much bigger, yet, IMO, has yet to find its place because its too 
big for single-user documentation generation and too small to be a 
content management system.

Moreover, nobody in the forrest community is interested in making 
forrest a full content management system.

So, in my vision, forrest starts from making static web sites and grows 
up, Lenya start with making supercomplex multi-user environments and 
goes down.

Will they meet in the middle? I don't know, but I can't see why not.

It might even be possible that leyna exists incubation by mergin with 
forrest, why not.

In short, I don't know the future, but all I know is that *at present* 
there is no community focused on serious content management and I want 
to fix this.

Why? because I need it for my future wild still-top-secret projects :)

>> We believe that the best architecture for CMS is the one described by 
>> Lenya where there is a clear separation between 'content editing', 
>> 'content repository' and 'content publishing', all of them left as 
>> 'framework' for developers to build customized CMS services upon.
> 
> 
> I reckon that. I would like to see it better articulated in that way in 
> the existing codebase, however.

Let me reply with the usual patent pending Jon's reply: 'patches are 
welcome' :)

>> We have most of the technology we need, we just need a community with 
>> a more detailed focus on these problems and so far the ASf doesn't 
>> have it.
>>
>>> * from what I read here [http://www.wyona.org/roadmap.html#1.0], 
>>> there is extensive refactoring planned _before_ reaching 1.0, yet 
>>> this is envisioned to be done as an incubating ASF (Cocoon sub-) 
>>> project. I am wondering whether it wouldn't be better if this 
>>> high-impact stuff is done before being moved to Apache (it also 
>>> sounds a bit like the Lenya people take the move to ASF as a given, 
>>> which might perhaps be a bit too premature).
>>
>>
>>
>> This is a good point.
> 
> 
> Pfew! :-)
> 
>> At the same time, I'd love to see refactoring done *after* entering 
>> apache because that would allow the community to steer the project 
>> before it's carved in stone (as it happened for xindice and it's now 
>> much harder to refactor)
> 
> 
> I have been lurking on Xindice since it was moved to Apache, and the 
> Xindice scenario is exactly what is inspiring me when reading this 
> proposal. There are some similarities IMHO.

True and acknowledged.

I'm *very* disappointed about what's happening on Xindice and it is 
partially my fault. I didn't understand that the original development 
community was data-oriented which doesn't make any sense *IMHO) for an 
xml database. Coming into apache created a shift toward 
document-oriented mindsets and the original developers left or found 
themselves uncomfortable with the new paradigm shift.

Xindice needs a technical leader and it's very hard to find one when the 
codebase is pretty solid and oriented toward a different set of needs.

>>> * reviewing the archived commit messages, I wonder whether the 
>>> proposed list of initial committers reflects reality, or that the 
>>> list has been expanded so that we won't have the suspicion it is 
>>> mainly one company/group working on the codebase (as is the case with 
>>> xreporter.cocoondev.org right now).
>>
>>
>>
>> Wow, that's very close to be insulting, Steven. Careful.
> 
> 
> It wasn't meant that way. It's a rough and possibly faulthy analysis, 
> while trying to verify what word-of-mouth told me. If I insulted anyone, 
> I'm sorry. But I very much value that criteria for becoming an ASF 
> project, so I think the question _should_ be asked.
> 
>> Anyway, I agree that the wyona community is not very diverse, yet, if 
>> Wyona had a successful and diverse community, we wouldn't pass thru
>> incubation.
> 
> 
> OK.
> 
>> But I wouldn't say "expanded" since all the people listed *did* in 
>> fact, work on the codebase and are interested in keep an eye on it 
>> (which is as good as you can get these days), the incubation stage 
>> will decide what happen.
>>
>>> * Xopus has recently had some troubles w.r.t. its licensing policy 
>>> (open, not open, etc...) Are these things effectively solved right now? 
>>
>>
>>
>> q42 has troubles understanding the oss business model. we are helping 
>> them privately. so far with very positive responses and some friction 
>> points we are trying to solve.
>>
>>> What part of Xopus will be inside Lenya CVS?
>>
>>
>>
>> None. If we reach an agreement with q42, Xopus will be submitted as 
>> another incubation proposal, apache licensed and they will be part of 
>> the community.
> 
> 
> Ah. Thanks for sharing! :-)
> 
>> In any case, lenya will keep the editing layer totally separated, yet 
>> design the proper hooks for them.
>>
>>> As I said, these are 'just remarks'. The fact I'm posting them means 
>>> I actually care about this proposal, in a positive sense.
>>
>>
>>
>> Hopefully my comments give you more insights.
> 
> 
> Sure thing. Hope I didn't offended anyone. I'm here because I care.

This is very appreciated.

-- 
Stefano Mazzocchi                               <stefano@apache.org>
    Pluralitas non est ponenda sine necessitate [William of Ockham]
--------------------------------------------------------------------



Mime
View raw message