Coccinella - Development

Mats Bengtsson 1959 - 2008

sander — Tue, 13 Jan 2009 22:30:19 +0000

Mats Bengtsson, the founder of Coccinella, passed away on November 29, 2008. He was born in 1959 and had a very successful career in academics and specifically in his field of interest, theoretical physics. Early in school he distinguished himself particularly in mathematics and he even produced his own math books in primary school when his teachers were not able to help him further on. He graduated at Lund university as an engineer, continued as a research student in theoretical physics and presented his doctoral thesis, ”Parton showers in a phenomenological context”, at the age of 28 years. He spent a post-doc period in Germany (Aachen), but previously he had also been at the Cern laboratory.

After his exams he started to work at ABB for a short period of time, before he ended up in Linköping where he received a research position at FOI, the Swedish Defence Research Agency. In Linköping he and his partner Mari stayed for quite a few years. During his entire life Mats had several hobbies such as gliding and model railroading. Unfortunately, Mats was struck by a chronic condition during the last 15 years of his life, but was able to work hard in periods on his special project, the Coccinella program which is an interactive communicative Internet program. He initiated and developed this computer program and released it as open-source on the Net so others may contribute to the program’s development and refinement over time. Hopefully Coccinella will remain and be continuously upgraded in Mats’ spirit and visions.

--
By Mari Lundberg, Mats girlfriend

Coccinella 0.96.12 *will* be released on March 18, 2009. However, this release will not introduce many changes because Mats was the driving force of the Coccinella project. Therefore, I will blog over the next few months about interesting projects and challenging features that Mats had on his TODO list. Hopefully, this will interest you to build further on the shoulders of a giant!

Story type:

Synchronized Releases and Greg Kroah-Hartman

sander — Mon, 03 Nov 2008 00:51:17 +0000

Credit: TiggerT, License: by-nc-nd

O'Reilly interviewed Greg Kroah-Hartman (log), the Linux kernel developer who founded the Linux Driver Project. The interview explains how the Linux Driver Project can support the weirdest Japanese USB devices, how Linux is being developed, why binary only kernel drivers are illegal, how a time-based release schedule benefits Linux, and how Greg thinks about synchronized releases. Although all these topics are interesting to read, I will restrict this blog post to the latter 2 of them.

Time-Based Release Schedule

Since 2.6, the Linux kernel has a time-based release schedule. As you can see, Greg really favours this development model (we also do):

It is working very well.
It is very predictable.
It increased our rate of developers, our number of companies participating, our number of developers participating.
There is better feedback.
We've got more people testing it out that way. People will only test when a release comes out.
The traditional "If it ain't broke, don't fix it" software engineering model is obsolete: the world changes, so we have to change in regards to that. Linus's famous quote is Linux is not intelligent design; it's evolution. We evolved the kernel.
Contributors plan their contributions: it's all volunteer-based. If you say "Here is when the timeframe is going to be" everybody knows that if they want to get something done they have to get it in by then. Otherwise, they'll work on it at their own pace.

Synchronized Releases

Even though we both agree that time-based release schedules are a good thing, Greg has a different idea about synchronized releases. Maybe I can change his mind by debunking some myths about synchronized releases:

I think Mark is asking other people to do his job for him. He has a very small number of developers; he has a very limited development staff. [..] If you look, wanting to rely on other people to stabilize that integration is a great thing. If other people can do your work for you that's a wonderful thing to get them to do; I don't think that's going to happen.

Myth 1: Mark invented synchronized releases to free ride on the community.

Wrong -- Mark did not invented this. The Eclipse Foundation did. Not because they wanted to free ride on the community, but because they wanted to spur adoption by providing more value for their users.
Myth 2: Mark wants this because he has not enough employees.

Wrong -- One of the issues of the open source community is that the same integration work needs to be done over and over again by different people in different projects. Synchronized releasing has the potential to increase the overall productivity of the whole community by reducing duplicate integration work. Hence, more can be done with the same number of people.
Myth 3: Nobody will ever synchronize their releases.

Wrong -- Gnome and Ubuntu are doing so, Eclipse is synchronizing (you really should read their goal!), we are doing synchronized releases at Coccinella, Midgard CMS is synchronized, and possibly others.

Trying to line up Red Hat and Novell and Ubuntu--just business model and internal business concern with each other is not fair to ask of a way of companies that compete against each other.

We have a very weird environment at Linux in the fact that there's these different companies that at the management level and the marketing level are competing. They're competitive companies. Red Hat and Novell go after the same customers, the same install base, and they're very, very competitive yet they have the exact same engineering staff. Red Hat is Novell's engineering staff and Novell is Red Hat's engineering staff. You can't do stuff to make that upset. It's a very, very weird model but just trying to align the business model and their timeframe of what they wish to do and trying to tell the Open Source community that you're going to line up with these businesses at the moment in time is a very unfair thing to ask.

Myth 4: Synchronized releases involve competitors which is not possible as they can't collaborate.

Wrong -- Strategic alliance is the business concept for companies working together. Yes, even competitors. There exists research about this.

Story type:

Jabber is Confusing

sander — Sat, 20 Sep 2008 02:41:29 +0000

Credit: nutmeg66, License: by-nc-nd

'Jabber' is confusing, not only for Aunt Tilly, but even for geeks:

Cisco announced today that it has acquired Jabber, an open-source IM and presence protocol used by Google Talk and Gizmo, for an undisclosed sum.

Cisco didn’t get ‘Jabber.org’ the open-source server on the XMPP protocol,[..].

Jabber is an application that we use throughout our office and is one helluva an IM.

Any news on keeping Jabber Open source?

Solution

Being devoted to a very friendly user experience in Coccinella, we decided a long time ago to progressively drop all references to 'Jabber'. For instance, whereas Coccinella formerly was described as a 'Jabber client', we do now talk about 'instant messaging client', optionally adding 'using the open standard XMPP protocol' for geeks. Even the JabberLib library is being rewritten from scratch. The long-term goal of this marketing decision is to kill the confusion barrier.

Story type:

Translate Coccinella 0.96.10

sander — Wed, 17 Sep 2008 07:43:15 +0000

Do you want to translate the upcoming Coccinella 0.96.10, planned for September 24th? Do you want to update any of the existing translations? Or do you just want to see which languages are ranking best? Then read on.

Translation status:

English (en): updated
Dutch (nl): updated
Spanish (es): updated
Italian (it): updated
Czech (cs): updated
German (de): updated
Polish (pl): updated
Swedish (sv): 162 new strings, translator needed!
French (fr): 533 new strings, translator needed!
Korean (ko): 665 new strings, translator needed!
Danish (da): 868 new strings, translator needed!
Russian (ru): 921 new strings, translator needed!

In case you want to update any of the above translations, please contact me, for example by adding a comment to this post. You best do not start right ahead to prevent 2 or more people updating the same translation.

In case you want to contribute a new translation, please download lang.msg.zip, read the translator instructions starting from step 4, translate the file, and send it to me. Of course, also this time it's a good idea to notify me in advance that you will do the translation. Remember that you don't have to translate all strings at once! You also can just start translating the most important parts of the interface, like menus, the setup assistant, login dialog, chat dialog, and so forth.

Coccinella will be so much easier with your translation!

Story type:

Tcl/Tk - Recent Advances

matben — Fri, 08 Aug 2008 12:04:25 +0000

Credit: freebird4, License: by-nc

There has been a lot of news in the Tcl/Tk community the last month or so, and I thought it's time to compile them here since Coccinella is written in Tcl. Tcl's windowing toolkit, Tk, has been "known" to be ugly and outdated. With the 8.5 release last December the tile package, now named ttk (Themed Tk), is included in the core which brings true native widgets on Windows (yes, Vista too) and Mac. Since Mac is my native platform, I can tell you that Tk beats both Qt and gtk, but perhaps I haven't seen the latest of them.

The question is then, where does this put Linux/Unix? There are a number of themes included in ttk which looks OK on these platforms, but there is nothing exactly like Qt or gtk. Instead, there have been a tileqt package written by Georgios Petasis which has delivered true Qt widgets. With ttk in the core, there have been some initial problems getting this to work due to missing stub tables etc., but it should work with the latest Tcl/Tk 8.5.3 release.

Just a few days ago the very same person announced tilegtk which does the same thing for the gtk toolkit. This is very, very, good news. Imagine that you can switch theme, and toolkits, on the fly without any program restart. This already works with the other themes and will likely work with tilegtk as well. My question to you is: which other toolkit has this flexibility?

I have already mentioned my recent advances with tkpath in this blog thread, and I will mention it again. This is the super canvas which is lacking in Tk, and which will bring not only tight SVG support, but also make Tk come on par with canvas like widgets from other toolkits. Still more to do.

Recently my attention turned to WebKit which has its history of its own. Dual licensed as LGPL/BSD it fits the Tcl community well. I have looked at some of the sources, and judging from the wxWidgets backend, it doesn't look terribly difficult to make it a Tk widget, but there could be show stoppers, like event loop integration etc.

When all this comes together it will make Tk a very competitive toolkit.

Story type:

SVG Graphics

matben — Thu, 17 Jul 2008 13:31:39 +0000

Just released is my tkpath package version 0.3.0 which for the first time brings a new canvas widget to Tcl/Tk which conforms to a more "modern" 2D drawing model found in SVG.

My motiviation for producing close to 43 thousand lines of C code for just this widget is because it is necessary for adding SVG graphics to the whiteboard in Coccinella. It doesn't claim to be a full featured SVG viewer. In fact, there is no support for translating svg tkpath yet, but since I have carefully designed the widget with SVG in mind, it shouldn't be too difficult to write such script code. SVG is really big and I don't claim every part is supported. For instance, clipping is yet unsupported. Remember, it is still version 0.3.0.

For the curious you can visit the SourceForge page where some screenshots can be found and the complete sources can be checked out. Before I have written a svg2tkp importer I don't have much fancy graphics to show.

Story type:

Synchronized Releases in Practice

sander — Sat, 05 Jul 2008 22:55:30 +0000

After reading The Blessings of Synchronized Releases, you probably wonder how we practically organize synchronized releases at the Coccinella project. In this article I will give you some insight into our release process.

Credit: Greg7, License: by-nc

How?

Synchronized releases are implemented at the Coccinella project this way:

Since 19 September 2007 we synchronize each 6 months with major Gnome releases.
Because only 2 releases each year is too few for a fast-moving instant messaging client like Coccinella, we also provide you with about 2 or 4 additional releases each year.
These additional releases in between the synchronized releases are not synchronized to anything.
However, we can say that there is a release roughly every 3 months. So, you can at least download 4 Coccinella releases each year.

Added Value?

You can choose:

Predictable synchronized releases every 6 months -- For non-business users, business users, and downstream distributions who would like to include Coccinella in their repositories.
Nonsynchronized Releases in between -- If you can't wait 6 months for new features in Coccinella.
Daily breakfast builds -- If you want to hunt for bugs in the newest development stuff.
Real-time development version -- For developers, hardcore bug hunters, and geeks who can't live without last-minute updates.

Story type:

The Blessings of Synchronized Releases

sander — Sun, 29 Jun 2008 22:11:54 +0000

People believe the idea of synchronized jumping to releases originates from Mark Shuttleworth.
People are wrong!

Credit: Xanboozled, License: by-nc-sa

Back in 2006, the Eclipse Foundation pioneered synchronized releases with Callisto:

The Eclipse Foundation is making history this week with what looks to be the largest ever synchronized open-source project release. The Callisto Simultaneous Release initiative coordinates 10 Eclipse project upgrades [..]

Only since 2007, about a year after Callisto, Shuttleworth started singing in tune. Nonetheless, the impact was greater as he continued practicing the art of synchronicity with much enthusiasm. Later, the choir grew when others started singing about the benefits of a regular release schedule, about cross project synchronisity, about release dates for Debian, about free software syncronicity, and about synching the open source release schedule.

Instead of repeating what has already been said in the above writings, I will focus on explaining the key blessings of synchronized releases. What are the huge potentials of synchronization for the broader open-source community? Why did we adopted the synchronization strategy? And why should your project consider adopting it as well?

Stevenotes and Windows Launch Events

What do Stevenotes and Windows launch events have in common? Yes, they both result into a flood of publicity. When Steve Jobs speaks, pregnant women's magazines write about it. When Microsoft spends a little money, the television journal will show us impressive fireworks that make New Year's Eve look like a non-event. All media eagerly report about these happenings!

The picture is very different for the open-source world. When a small open-source projects such as Coccinella issues a new release, you may at best see a review by a blogger or by some obscure local computer magazine. It gets slightly better with bigger projects which may attract the attention of several bloggers and a dozen computer related news sites. Even more, the real big projects may sometimes even go a bit further. However, no open-source project approximates what Apple and Microsoft do. No pregnant women's magazine writes about Firefox 3 and the television journal will not be interrupted for a boring interview with some geeks about Ubuntu's Intrepid Ibex.

How can we copy the marketing successes of Apple and Microsoft to open-source? Some may suggest we need someone like Steve Jobs, whilst others would like to spend more money on launch events. They are both wrong! Trying to copy Steve Jobs or Microsoft's deep pockets is mission impossible. If Apple or Microsoft notices we are trying to catch up, they will simply improve their key strategic advantages and catching up will become even more challenging for us. What do we have to do then?

Community is in what we excel; our key strategic advantage. To repeat the marketing success of Apple and Microsoft we need to leverage the community. Massive synchronized releasing is the perfect tool to achieve this leverage. By synchronizing, release related marketing efforts of multiple communities are focused at the same moment and can leverage each other. The rhythm of so much projects doing their marketing efforts at the same moment will shake the Internet. I'm sure the pregnant women's magazine and the television journal will feel these shakes...

Yes, massive synchronized release cycles have the potential to create a bigger impact than Stevenotes and Windows launch events.

Planning Drupal Upgrade

This website is still powered by the old Drupal 5. Even though its successor has already been released months ago, we cannot tell you when exactly the Coccinella website will be upgraded to version 6. If you are also running a Drupal based website you should have correctly guessed that our website depends on third-party modules. Most of these modules did not had an updated version available at the time of the 6.0 release. Even today there are still plenty of them that have no stable 6.x version or that have no version at all for 6.x! Of course we can look around for alternative modules or we can just drop website functionality, but obviously, both will not contribute to a painless upgrade experience.

I chose the above example because it clearly shows you that synchronized releases also benefit non-business users and not only business users. If Drupal and its third-party modules had adopted a synchronized release scheme, we could have switched shortly after the 6.0 release, without the need to lose time on looking for alternative modules, and with lower upgrade risks. It would also have been possible for us to plan the website upgrade months before the actual release of Drupal 6. For businesses this translates into 3 cost savings: labor cost savings, risk protection savings, and savings thanks to the possibility to plan the upgrade a long time in advance (e.g. hiring additional employees to assist with the deployment is cheaper when it is done a long time in advance than when it is done only a few days before the deployment). Hence, synchronized releasing adds value for both business and non-business users, upgrading open-source software will be easier and cheaper.

In the long run, we may see an additional blessing of massive synchronized releases in the planning area. When choosing a release strategy for their software, companies who provide cross-platform software may decide to align their release processes to the platform with the best predictable release dates. Accordingly, commercial applications, games, and hardware drivers may closely follow the synchronized release cycle of the open-source world. Compatibility between commercial software and new versions of open-source software thus could be greater than compatibility between commercial software and new versions of proprietary software.

Release-Time Collaboration

When open-source projects synchronize their releases, they all will at any time be in a similar development stage and thus their priorities will better match. It should be clear that collaboration among projects is easier when all noses are pointed in the same direction. Projects will face the same problems and can better understand the needs of other projects. Hence, discussions can be shorter and projects can move forward sooner. More time can be used for coding and thus productivity will increase. Open-source projects will be able to provide their users with more new features, they will be able to fix more bugs, and code freezes can be shorter without losing software quality.

What's Next?

Next week I will write about how we practically implemented synchronized releases at Coccinella. As our experience will show you, adopting synchronized releases does not mean your project loses its freedom. There are multiple ways to implement the synchronization strategy.

Story type:

Tcl/Tk 8.5

matben — Tue, 27 May 2008 14:40:49 +0000

I have just removed the compatibility code for Tcl/Tk 8.4 in cvs, and Coccinella therefore now requires Tcl/Tk 8.5. Since 8.5 was released already in December last year I thought that five months should be enough to allow for people running Coccinella from sources to get a Tcl/Tk 8.5 installation.

Well, there is not much to install if you run a single executable tclkit which can be found for most platforms. If you don't find your platform there it is not difficult to build Tcl/Tk yourself. Since the tile package is contained in the core Tk of 8.5, only the treectrl and tkpng packages require additional builds.

Speaking of 8.5, the direct reason the switch is made now is that I used the dict command for the new themeing engine. As typical with Tcl, commands (and manual pages) look deceptively simple. The hardcore programmer who is used to pointers and reference counts can't understand that Tcl has it all built in. Dictionaries can be seen as a hash table that maps from arbitrary strings to arbitrary strings. But here it can also contain other dicts in a recursive manner. This makes it very useful for representing a kind of arrays with multiple keys. Before 8.5, the standard way was to form a string like arrayName($key1,$key2) where "$key1,$key2" was used to lookup the value. This was clearly very ugly. With dicts you set values as:

dict set myDict $key1 $key2 ... $value

which implicitly creates a hierarchy of dictionaries.

Story type: