Hi

So let me tell it the way it is. Most of the time we're developing against our "HEAD" code line - that's the one that everyone sees live in SVN. Any bug that's found and fixed goes straight into that code line. In the past, whenever we've got to the end of a release cycle, we create an Enterprise code line and do additional intensive testing and fixing. All fixes made in this code line are also merged to the HEAD code - sometimes immediately (always if a fix is provided by the Community), sometimes once we've got back into the next development cycle.

The Enterprise code line then continues as a stable code line for maintenance releases. People paying for support will raise issues and expect them to get fixed, which we do on this stable code line. These fixes also get merged across to HEAD.

There are nightly builds of HEAD made available from our Download page, so you can always grab the latest without having to build it yourself. For your case, you're going to have the fixes you crave. We will start to do more regular Community "releases", but we need to work out how best to manage it (we don't want to divert too much effort from development).

What this all means is that HEAD is always on the bleeding edge, while Enterprise code is more reliable and predictable. If someone's going to use Alfresco in production, then they're devaluing their business by skimping on paying for support. Fortunately, most sensible companies are paying for support and stability, so Alfresco has been able to grow strongly. Building Enterprise software isn't a hobby, it's a commitment that needs clever and experienced people.

So, take a look at Nutsio, that's how this Open Source thing is supposed to work. See where they've taken design ideas from us - who do you want to be with, the leader or the Alfresco wannabe? Make up your own mind [but take a look at the number of Documentum partners that are doing Alfresco implementations, and bear in mind the places where we're displacing Documentum and Interwoven ;-)]

And remember, *all* fixes we make are made available to the Community. We've made huge leaps of faith to keep our Community with us (e.g. going GPL). And the hours our core developers spend on the forums each day answering questions is exceptional. Take a look at Adobe's announcement about Open Sourcing Flex - why is Alfresco mentioned? - it's because we helped persuade them. Questioning our openess and commitment to Open Source is unfair.

Cheers
Paul.

Which brings up another point ... Alfresco has a track record of introspection and change. It's demonstrable. The company isn't open source because it's a sales gimmick, they are open source because it works. They have tried, learned, changed, tried again and on and on. In short, Alfresco is alive and it is evolving as it grows and as its environment changes. That is something you won't find in every open source company or even in most companies. Alfresco is still asking the questions today not acting on yesterday’s answers.

So are these bug fixes available only in HEAD? If so, as a practical matter, how would one separate these out to be applied to a stable branch, if one didn't want to track HEAD?

Having just written that, effectively what we'd be doing is tracking the 'Enterprise' branch, so I suppose the question is:

Why not open up the Enterprise branch?

This isn't necessarily to the detriment of Enterprise customers -- their bugs and priorities are the ones that determine the fixes that go into that branch.

Regards,
- SteveN

It isn't.

But it's not really just the bug fixes that are the problem. The current
situation means that potential contributors are instead wasting their time
working around issues because they can't see the source they need
'through the keyhole', and so are actively discouraged from committing time and effort to Alfresco.

I don't think that's actually true, Russ, but even if that were currently the case:

as long as it reamins the case it means that Alfresco has failed to build a
meaningful community around its product!

increasingly, the non-availability of the code for any given build is
becoming a deal breaker (I've seen this more than once).
People who buy into open source want, surprise!, open source.
Open-ish just doesn't cut it! I've seen plenty of Microsoft's
source code, but this doesn't help me if my PC isn't working,
particularly if the version I've seen isn't the one causing my problem

The talk of "doing community support all the time and that costs
money" and having to "keep the lights on and pay bills and
salaries" simply underlines the fact that the 'Community' is
currently seen as a burden rather than an asset (or perhaps as a
markeing asset that comes at a price), and as long as it is
regarded and/or treated in that way, it may very well remain so.

You may feel, Paul, that you have "made huge leaps of faith to keep our
Community with us", but the point here is that Alfresco hasn't really
made that leap of faith yet.

Don't misunderstand, I'm not questioning the fact that Alfresco wants
to do Open Source, and to do it well. However it's clear to me that, in
spite of having some high profile OS names on board, the company
as an entity hasn't really come to terms with OS yet.

There's a lot more to be said, but it's too late tonight and I'm too tired...

Thanks so much for making me aware of this competing product (Nuxeo, though it is much less mature). I was unaware that any other ECM existed that offered the capabilities (at least the ones i care about) that Alfresco has.

After seeing the fire storm of complaints on Slashdot about whether Alfresco really honors the meaning of opensource it made me realize the troubling reality that due to the non transparent way that they manage there subversion repositories and the (expected) unstable state of there HEAD branch, the prospect of getting a more stable build with current bug fixes is very unlikely.

I also see that they have some interesting features on the roadmap (v5.5) such as 21CFR Part 11 support via the (out-of-the-box) ability to digitally sign electronic document inside a workflow (approve/sign ; reject/sign). A feature that myself and others that work in heavily regulated industries (like Pharma) need and cannot imagine deploying any solution that lacks such a feature (see my post on minimum requirement of 21cfr part11....awareness of person signing doc ).

I would like to say that i am impressed with Alfresco and it is the most feature rich (also has far more developers). However there is a growing sentiment that it is not truly operating like an open source company with regards to code access and transparency and it is something that i believe will cause people to migrate to other solutions even less mature than Alfresco.

Come on, this is a bit cheeky - that sounds like an advert for Nutseo on our forums. Good luck with it, we have *no* worries about them. We're completely open about our model and it is open source.

Paul.

Paul et al,

I do have a question about fixes being provided and how they are distributed. If fixes are happening in the Enterprise fork of the product and these same issues may be corrected by someone in the community, how are you resolving these two different fixes? It seems that the community isn't really being asked to participate in the product because if I do make any changes to source, they could be undermined by an 'official' fix that moves from Enterprise to Community at the next release date. Am I wrong?

Alfresco has full visibility into all commits (to both the labs and enterprise branches) so this is unlikely to become an issue in practice. On top of this, bug fixes are merged into the labs branch soon* after being fixed in enterprise, so the window between a fix being made and it appearing in the labs branch is usually small (typically of the order of days, although as with anything there are outlyers), further reducing the chances of this kind of conflict occurring in practice.

* where "soon" is defined to be "as soon as is possible, given the the engineers' workload" - it's in no one's interest for the labs and enterprise code bases to diverge too far, so there's a natural incentive to merge bug fixes across as quickly as circumstances allow.

Cheers,
Peter

It is worth making one thing very clear: if someone from the Community contributes a fix or a feature, that goes into the Community code first. Always.

We don't hide what fixes have been made in the Enterprise maintenance releases and they always get into the Community code. However, in the real world, where we are paying full-time, world-class developers we obviously have to concentrate on the commitments we make to people that pay for our Enterprise support and maintenance. If we don't, there will be no continually growing Alfresco. Look at the speed of development of most open source projects that are done by people who are not on it full-time and you'll see that it's usually glacial. You want to have your cake and eat it (http://en.wikipedia.org/wiki/Have_one%27s_cake_and_eat_it_too) for your "true" open source.

I think you'll see with the new application and framework that is coming with 3.0 that there will be much easier ways for people to contribute functionality and components without being Java gurus. The easier it is for more people to implement extensions, the more will become available to the Community.

Paul.

In retrospect i believe that it was wrong to make such grandiose statements about Nuxeo, frankly i have not used their product and i am sure it has its own ills. I did however state/imply that Alfresco is much more mature and I also (at the risk of sounding overly sycophantic) I believe it has an excellent engineering team.

My company likes Alfresco and while our initial pilot deployment was marred with problems (due to the instability of the HEAD branch code, the only branch available) we liked what we were able to do in/with the product, except for the obvious lack of support for 21CFR Part 11 style sign/approve and sign/reject and the absence of this feature on the published Roadmap (if/when we do purchase Alfresco there will have to be guarantees that 21CFR Part 11 is fully supported).

So maybe my not too subtle attempt at bringing to light my desires has only served to get the attention and raise the ire of a developer, but I would think that these are important issues being raised in this thread (also on slashdot) and while it maybe fine to say "feel free to move on to competitor X" I would think that a better approach would be to see if there are concerns that are of merit in the community that are worth investigating regarding licensing (ok, in my case features also).
I recently had this discussion with a senior member of the Compiere product group about the lack of product/documentation openness (unlike Alfresco there was no active "community") and this resulted in users migrating to other solutions and a fracturing of the community/product, they have since adopted a more open/transparent process.

The truth is, with the nature and complexity of ERPs (in the case of Compiere) and ECMs (i.e. holding valuable customer data) a user/company cannot afford downtime or outages and for that matter data-loss. These are products of information aggregation and all but the small mom and pop businesses will think twice about deploying these platforms without support contract.
So my argument is that Alfresco Inc. should listen to the concerns/rants and if possible offer solutions that are long term beneficial to the company and the testers/users (of the opensource version).

-Kurt
P.s. Also add 21CFR-P11 support to the roadmap !

Can I also add that while I see a lot of mention about access to code but not about stability releases

Can I refer you to
http://forums.alfresco.com/en/viewtopic.php?f=14&t=14330&p=47380&hilit=Stable+Version#p47380

Now I really like the essence of that post - community support is going to get better - but it clearly is not good at the moment - apparently there has not been a "stable" release since 2.1 and there have not been any patches to the 2.1 release.

Supporting paying customers understandably takes a priority - but there are bugs in 2.1 that are causing me trouble and I can vouch for the instability of the more recent community releases.

That would be unfortunate. It would also be foolish, IMHO. There are two good reasons that Nuxeo is not as good as Alfresco, and likely never will be:

1. It doesn't have engineers with as much talent and experience in ECM as Alfresco has.
2. It doesn't have a business model that will allow it to scale to the demanding enterprise requirements for which companies use Alfresco.

I understand that you would like to have Alfresco give you 100% of its code, 100% of the time, for 100% no cost. Believe me, I'd love the same thing from Apple: my Mac for free, plus all the software for free.

But don't you think it's fair to pay for value you receive from Alfresco? Let's face it: you are unlikely to contribute any code back to Alfresco. You're also clearly unlikely to contribute any cash back to Alfresco (or Nuxeo, for that matter). Why is it wrong for a company to seek to find ways to give ever-increasing amounts of code away for free, but also find ways to charge prospective customers for value received?