How not to destroy your Open Source Community before it even exists.
Published by admin on Monday, June 7, 2010 - 00:08:42 - Filed under General, News, Announcements, Development
While being enthusiastic about Servoy’s announce that Servoy 5.2 will be Open Sourced (see my post on that subject), I really hope that they are going to do it right. What do I mean?
Most major Open Source software are backed up by companies who have interests in it (think of Eclipse and look at how IBM/Adobe and others are major contributors to the project), so it is not unusual for a commercial company to Open Source a product, or even pay part of its staff to maintain one. I will not list again the many benefits that this move can offer (just see what Jan Aleman said in the official press release), the problem is that there can also be very bad consequences if the company is not “open” enough.
About openness, you might remember the famous post about the failephant of the Ushahidi project. This is something to keep in mind.
Yes, in matters of Open Source, there is a flip side of the coin: benefits on one side, big dangers on the other. What will make a big difference when you draw is the company’s attitude towards openness in general.
Now since I like the platform very much, I would like Servoy to get all the benefits and none of the drawbacks.
Josh Berkus, CEO of PostgreSQL Experts Inc (quoted “one of the foremost PostgreSQL consulting firms in the world” on Servoy blog), lately summarized how easy it was to destroy an Open Source community.
Interestingly enough, the Josh Berkus who did this presentation is the very same guy who will be doing next thursday’s webinar about Open Source Servoy and PostgreSQL…
Please read his very informative (yet funny) slides called “10 Ways to Destroy Your Community” - there is also an abridged version for the impatient (funny and informative as well), called “Destroy your Community in 5 easy steps“.
For those of you who also want some background comments about these slides, you can read the post Josh made about Sun and the Ten Ways.
If Sun and dozen others did it so badly, there is still hope that Servoy will do it right…
For now, I would particularly insist on point 3 of the “10 ways”, point 6 is very important of course, point 7 and 9 seem obvisouly also very important… Anyway, any of these points are important.
Now you would understand when I say that there are dangers if you don’t do it right, so I hope that Servoy will especially refer to the last slide of the presentation and keep that in mind… and act accordingly.
Because in the end, and Servoy needs to be very careful about this, if you don’t do it right, it’s worse than not doing it at all!
Let’s just hope that Servoy will be listening to Josh Berkus good advices and not fall into the trap of destroying its Open Source Community before it even exists.
Personally, I cross my fingers…
Monday, June 7, 2010 - 21:15:29
Very true Patrick.
It’s easier to kill off your community than to build one up.
Perhaps Servoy should also read the book The Art of Community from Jono Bacon. The book itself is available for sale but also as free download.
http://www.artofcommunityonline.org/2009/09/18/the-art-of-community-now-available-for-free-download
Anything to make this transition to OSS a success!
Monday, June 7, 2010 - 21:55:11
An essential book indeed Robert!
Thanks for sharing!
I do hope that Servoy is going to read it.
Monday, June 7, 2010 - 23:50:33
Yeah, open source can be difficult - if you’re not prepared as an organization to handle what being an open source project owner means… I have faith that the Servoy folks have thought about this and have (hopefully) planned to avoid making mistakes.
However, I, for one, would encourage all developers with extensive experience in open source to help Servoy and their liaison(s) to make this transition as smooth and error-free as possible.
Good article, Patrick!
Tuesday, June 8, 2010 - 00:46:33
Thanks Bob!
you are right, Servoy will need all the help they can find, which is why I wrote this article!
Be sure that I will do everything I can to help making the transition.
Tuesday, June 8, 2010 - 03:51:24
Good and instructive article, Patrick.
We’ve used Servoy platform and we admit that it is a good and rich one.
I want to advice Servoy community that in my opinion, Servoy documentation is unfortunaltely not at the level we expect!!
Tuesday, June 8, 2010 - 04:07:56
Without the Server being open source the stated goals of this “open source” project by Servoy is dead on arrival.
There is going to be no stampede of developers rushing to check Servoy out (I’ll run my numbers in several months and prove it), tons of bugs won’t suddenly get fixed, features won’t be rolled out nice and polished, documentation won’t suddenly improve, etc etc.
Servoy could have accomplished the same and much more by just opening up their bug system, give partner level SAN members access to source code, and then spent the rest of their time on developer resources (like documentation and better example files).
Instead, now they have to manage an open source project now in addition to managing their end users. Which means additional documentation, better community management, and more transparency. All of which Servoy has proven to basically suck at. Why would these skills change all of a sudden because something becomes open source? Who’s going to spend time on the average Servoy developer who just wants to code a solution now?
It stinks of a PR stunt. Nothing has changed. Oh wait, we’re promised a bug tracking system. Well that’s cool — except it won’t include the 3rd and arguably the most important piece of the puzzle — Servoy server.
I would love to be excited but I can’t help but be completely cynical. The open source webinar today was joke. Nothing was polished, no community systems in place, no clear direction on how to do things and what the benefits are. Just a ton of marketing speak, promises to improve (which we’ve been hearing for months), and the most basic demo that you can possibly get away with. Mirror image to the release of Servoy 5 (which was total disaster).
If this is how Servoy goes open source, Servoy is going to be a case study on how not to do it in the next version of “Art of Community” (a very good book).
PS: And can all the Servoy people just quit with the “The world’s most advanced open source database” line? When asked a direct question, repeating a PR quote 5 times is retarded.
Tuesday, June 8, 2010 - 19:24:37
All,
Howdy. Yes, I’m working with Servoy on open-sourcing their product. However, I just started talking to them a few weeks ago. Feel free to contact me if you have questions or concerns which you think I could address. Of course, I am primarily involved on the PostgreSQL side.
David,
You raise a number of issues of which I was previously unaware. You’re certainly correct that having a new license doesn’t make other chronic issues go away. Do you think that cultivating a contributor community would help address some of these shortcomings, or do you think that would be unsuccessful?
Tuesday, June 8, 2010 - 21:08:17
Servoy 70% open source vs. Servoy 100% open source is two completely different things.
In my opinion, Servoy marketing and message around their open source push is touting what the advantages would be if Servoy were 100% open source. Then in yesterday’s webinar “the other shoe” dropped.
It has been a point of contention for sometime now that Servoy is releasing features in their product that are more marketing driven than actually useful. Servoy v5 was released with tons of “features” — many of which were incomplete, barely out of alpha stage, buggy, or at the very least not quite polished. Additionally, there was virtually no developer materials on all of the new features. As a result, there was a huge expectation gap between marketing and reality that has caused a significant headache among Servoy developers.
This doesn’t mean I don’t like Servoy 5. It’s an amazing product. But it has flaws and Servoy is/was very coy about admitting to the flaws. If they had been more upfront with the limitations to begin with, developers wouldn’t have been near as frustrated.
I see the same thing repeating with this open source initiative. I think there is huge value in the 70% open source route:
- open bug tracking (finally)
- easier for plugin developers to write for servoy
- more direct access to servoy engineers and sharing of ideas
But it is definitely not 100% open source and should not be spinned as such. Servoy still has a lock on the most important piece of the software (the server) and the open source pieces don’t run without it.
As a result, can we really call this Servoy “going open source”?
I certainly don’t believe that one of the stated goals — “the vast expansion of the Servoy eco-space” — is realistic under this pseudo-70%-open source-whatever Servoy has cooked up.
But if it results in Servoy becoming more transparent (and it seems from background discussions this is true) then I’m all for it.
Tuesday, June 8, 2010 - 21:38:47
David, you are absolutely right. Now you are talking business! It’s a tuff but a fair assessment of the issue.
I think that this SOS project is in part the result of the push that Patrick, Roberts, Jeff, David and others gave to the company to move toward a more open environment. Undoubtedly, Servoy is moving toward that. As described by David, Servoy is moving toward (1) opening their bug tracking/fixing system, (2) making easier for developer to write plug-ins and (3) creating a new channel for more direct access to Servoy engineers and sharing of ideas. This however, in my opinion as well, it’s NOT open source.
Josh, about the question you ask to David re cultivating a contributor community, my answer will be YES. The Servoy community has always been very supportive of the product and the company as well.
We are drafting at the moment ideas about how to start an open source community around what Servoy is offering at the moment. (This is evidence of a community that wants to contribute). Do you have any other ideas in addition of what is Servoy already doing? How can we effectively have a community of contributor that can actually contribute? What channels is Servoy putting in place for this?
Tuesday, June 8, 2010 - 22:00:30
Just FYI: I just learned from Jan Aleman that there are many OS projects (such as MySQL, JasperSoft, etc.) that do not have 100% of the code open source. Why not Servoy then? This is a pretty good point…
Tuesday, June 8, 2010 - 22:13:28
Juan-Carlos,
While it’s certainly true that there are many other OS projects that aren’t 100% open source it’s optional, value-added components that remain closed. In this case though, the core of Servoy is not being open-sourced so I’d have to agree this is more of a marketing thing than something of real substance.
Corey
Wednesday, June 9, 2010 - 00:00:21
Thank you Corey. I’m learning from this ordeal
Wednesday, June 9, 2010 - 01:16:04
The core of Servoy at developmentime is Developer, the core at runtime the Client, they form 94% of our source code.
It would be insightful yeroc if you can share what you cannot get done with what you now have access to
Wednesday, June 9, 2010 - 01:48:58
Jan,
You have to consider the server component a core component because without it you can not run even the most trivial solution with Smart/Web Client nor even launch Developer. This is unlike the two examples given above where you can certainly use MySql or JasperReports with strictly open sourced code.
In my case I wanted to explore the possibility of having a FoundSet backed by an IDataSet on the server that kept the ResultSet open for at least a short period of time allowing for very fast forward iteration over a FoundSet without 100s of queries being generated in the background as happens today. Whether this is feasible or not with the current architecture I do not know although on the surface I would think it is doable. Right now though, it’s impossible to pursue as this would have to be implemented on the server.
Corey
Wednesday, June 9, 2010 - 16:45:01
Are there any other open source projects where the owner charges for their product that is open source? Or 94% open source?
Servoy is still a commercial product where it counts: you have to buy it (at a significant cost) to do anything useful with it.
This is the main reason why this seems like a marketing gimmick to me. It’s too easy to think that Servoy is trying to get all the advantages of open source while still retaining their current business model — which is to sell user licenses.
There is a community edition of Servoy that has been out for a couple of years and is free for 5 users. If Servoy were to bump this up to say 100 users as part of this open source move, I think that would go a long ways to balance out the perception that Servoy is doing this just to gain marketing exposure.
I think it would also be the key to retaining new Java developers checking out Servoy for the first time. Which really should be the primary point of this open source move.
Wednesday, June 9, 2010 - 17:24:36
David,
There are a lot of products/libraries like that. Usually the product is dual-licensed in that case. Typically the GPL/AGPL license is chosen for the free version which means if you’re building a product based on the source-code you would have to open source your product or alternatively pay for the product under a more commercial-friendly license that doesn’t force you to open source your code as well.
I have absolutely no problem with Servoy charging for the product. Moving the limit from 5 users to 100 users would drop Servoy’s revenue drastically. I don’t see how the company would survive that way, do you?
Corey
Wednesday, June 9, 2010 - 18:54:24
I see your point — and I agree Servoy needs a way to make money.
How does it work then if I want to release a solution built with Servoy as open source currently? (For example, a CMS.) Would the end user need to buy Servoy licenses in this case?
Wednesday, June 9, 2010 - 19:54:17
David,
As far as I can tell nothing changes in terms of what it costs to develop or deploy a Servoy solution. The only change is that some of the source code is now available. As pointed out above without all of the source code being released it’s of limited value to outside Java developers. On the other hand, if Servoy open sources the core Server component then it may become difficult to charge money for the product (though presumably most people would be willing to continue to pay for support). I assume this is the reason why the Server itself wasn’t open sourced as well. Additionally, there might be a risk (though probably low) that someone else would take the source code and create a new product called Yovres or something like that. On the other hand that particular risk is mitigated by the license they chose which would force any such competing products to be open source as well.
Corey
Thursday, June 10, 2010 - 00:09:50
There are several advantages to Servoy making their product open source:
1) To be a developer product in today’s market, you pretty much have to be either open source, or be Microsoft, Oracle or Apple.
2) As with MySQL, users have some assurance that they will be able to continue to use open-source Servoy, and fix bugs in it, even if Servoy Inc. is acquired by a company with a different business model for the product.
3) Users have the ability to fix bugs which affect them without waiting for a new release of Servoy.
While I agree that the fact that Server is still closed limits the above somewhat, as I understand it the plan is to eventually open source everything. However, just as with other companies, Servoy can’t necessarily open source everything with the flip of a switch; they have third party licensing and existing customer agreements. Not to mention changing the business model. These things take time.
More importantly, Servoy needs to go from a company with a Filemaker-like closed culture to a more open and online culture. In the other companies I’ve worked with, that transition takes years. So, please have some patience with that.
I am please to see so much interest from the Servoy online community in the OSS release.
Thursday, June 10, 2010 - 01:17:39
Thank you all for your comments!
I guess we all agree that Servoy Open Source is a great idea, but that right now it’s just the start of the idea. Of course, Servoy needs to find a way to sustain its revenue but to be able to release the server part as well so that the transition is complete the licensing model will probably need to change completely along the way.
The idea of this post and most of these comments were only to state that the community is eager for these changes and on the look-out about what it can bring (good or bad). We care that Servoy will not fall into bad the bad traps that Josh enumerated because it would render the process totally useless and even potentially detrimental.
Servoy lately reached out to the community and there are some discussions involving long time members where a few things have been clarified and agreed on already. So there’s hope!
But there’s a lot to do yet, and indeed we will need a lot of patience.
PS: Welcome Josh to the Servoy microcosm! It’s little yet but as you can see, it’s full of passionnate people who share the same goal, so a good basis for a great Open Source community.
Thursday, June 10, 2010 - 01:25:00
Welcome to the world of PostgreSQL as well!
The other part of this announcement was that Servoy is going to be shipping with PostgreSQL 8.4 instead of Sybase ASE. We’re hoping this means that Servoy users will get involved in the PostgreSQL community as well. Certainly that’s how I got involved with Servoy.
Thursday, June 10, 2010 - 01:48:41
Thanks Josh!
We’re looking forward to get involved in the PostgreSQL community too!
Thursday, June 10, 2010 - 21:48:44
@Yeroc,
I would suggest you post your thoughts on the idea’s about the cursor-backed FoundSet implementation on the Servoy Forum, to see what people think and get some feedback on it’s feasibility.
I might be mistaken, but I think this might very well be doable: The application server is not open sourced, but large part of the code that is open sourced runs serverside.
Paul