FOSDEM Preview
The Document Foundation was announced on September 28, 2010. So far, it has been an umbelievable ride, especially under the development point of view. Our core development team has managed to attract close to 400 new developers, and has achieved a large number of the ambitious goals set on that date. We still have quite a long way to go, but LibreOffice 3.5 – due next week – will be the very first release showing TDF “development directions” not only to geeks but also to end users: a leaner and cleaner office suite, packed with new features. If you happen to be in Brussels for FOSDEM, you are warmly invited to join our DevRoom in Building H or walk by our booth on the first level of Building K.
You can download a PDF or a JPG of the infographic, for printing or publishing on your website or blog. From now on, we will update it on a monthly basis, adding more numbers as soon as they will be available.
Are there any stats on how much work is devoted to each component application of LibreOffice people (Writer, Calc, Impress, …)?
We have published all the reliable statistics we have, and I think that we have already done something quite unique. In addition, splitting commits by application would be not only difficult, but could also lead to wrong assumptions, as – for instance – some applications rely more on the graphic engine, but the improvements to this component would not be related to the application itself.
Could you share with us some info about the tools used to count the commits etc (e.g do you use software that counts commits from Git)? Is the procedure automated?
We use different sources: Ohloh, Git and Bugzilla. We pick from Ohloh the number of committers and commits (histograms), from Git the numbers related to developers and their affiliation (pie charts), and from Bugzilla the numbers related to QA (bar and pie charts). I get the numbers from Ohloh (the easiest task), Michael Meeks gets the numbers from Git (I suppose with some kind of script), and Rainer Bielefeld gets the numbers from Bugzilla (again, with some kind of script). I then chose the data and create the visual representation.
Good stuff. Let us know if the scrips are available anywhere (of become in the future).
Thanks for your response
—Kostas
Wow !! … amazing infography. Thanks a lot Italo.
If you are going to update it every single month … then some stats (‘Hackers by Employer’ pie for instance) could show the monthly evolution instead of a ‘total pie’ … I mean … it’s just an idea …
Off-topic .- According to gaining volunteer development traction (or maybe consolidating some already gained new volunteers) … I wonder if this season’s bet on GSOC will be stronger than previous year.
Thank you.
What software was used to create the graphics? They’re amazing.
Guessing should have been very easy, as everything has been created – of course – with LibreOffice. Each graph has been created with Calc and pasted into Draw for some visual tweaking, and then assembled with Draw to build the infographic. You do not need another tool when you already have a great free software !!!
we complimenti per il lavoro che state facendo! spero che in futuro migliorerete funzioni anche il dizionario(quello di office è migliore al momento)… e anche la visualizzazione delle pagine e il caricamento delle stesse. Con libre office 3.4 per aprire un documento .doc da 5 mb ci ha messo 20 secondi, al contrario office2010 ce ne ha messi 2-3 per aprire il documento e intanto precaricava in modo che già potessi utilizzare word. (quindi spero che introduciate anche il precaricamento).
comunque sono sicuro che se continuate cosi presto sostiruirò office,ma al momento c’è tanto lavoro da fare e mi trovo ancora meglio col primo.
Great infographic.
May LO grow from strength to strength!
Really impressive to know that all graphics have been created only with LibreOffice. Sure I will try your suite and will post my thoughts a couple of weeks later.
I do not think it is impressive, it is business as usual…
When will there be a open SaaS or open Cloud version of LibreOffice? We will need a open source competitor for Office365 en Google Apps !!
Very useful and clear document! Thanks to all of You! 🙂
There are many people (frequently not reporter) that add interesting and useful comments to various bug report so that another useful histogram may be one that show #commenters over some #bugs classes (1, 2-3, …) similarly to “Volunteer reporting a total of bugs”. Moreover with the option of excluding reporters’ comments may yield an interesting statistics about users use of, and confidence in, bugzilla and in general of their help in solving bugs.
What do you think about this?
Carlo
If the document is only available as PDF or JPG, does that mean that it has not been done completely from LibreOffice itself?
If it is in fact an ODF document it would be helpful to have it has a showcase of the suite.
The document has been done completely with LibreOffice, but I am slightly reluctant to share the ODG because it could be easily tweaked. Let’s find a creative solution to share the document without incurring in this kind of risks.
Next time please include QA team (guys who transforms UNCONFIRMED bugs to NEW or INVALID but reports bugs rarely). We are working too.
I am happy to include whatever data I get, if someone gives me the data (which I am not able to get by myself). I know there are many more people working around LibreOffice, but I have to rely on data extracted by someone else.
There is a discussion about this on the LO-QA mailing list.
Probably this information willbe available at http://www.libreoffice.org/about-us/credits/
The charts look nice, but they are completly useless to get the really important information out if them:
How are these numbers releated to a) the old OOo dev process and to b) the parallel dev process of Apache OOo? Are they higher, lower, equal? How many new features did every fork got in this time and how large were the features?
THIS are the questions everyone is interested about.
We are not going to comment about Apache OOo in our premises, now and in the future. Of course, we have the numbers, which are quite easy to find on Ohloh (and I have blogged about them on my personal blog, which is quite easy to find with Google). In general, the number of developers at LibreOffice is higher than the number of developers at OOo (although the number of full time paid developers is slightly smaller) and at Apache OOo (we have more full time paid developers and far far far more volunteers).
The comparison on new features is difficult, because we have added features incrementally in three different major releases (the third being out this week), while Apache OOo has spent most of the energies in getting rid of copyleft code and have added only one feature (but I might be wrong) which we are free to integrate as all features developed for Apache OOo (while the opposite is not true, as we release code under LGPLv3+/MPL and Apache cannot integrate that code because of the copyleft license).
Thanks for the hint about your blog. There we are really getting to the more interesting answers. 😉 Are you publishing these comparing stats frequently?
Comparing absolute commit counts is to be taken very carefully, however. These numbers don’t say anything about the commit content – which is actual important thing about programming. Let’s say LO implements a feature in 50 single commits, while AOOo does the same feature in 1 single commit – who has done “more” work? This counts in both directions, of course.
This has also to be taken in consideration for a real significant and meaningful statistic.
The idea is to update stats on a monthly basis.
I absolutely second your comment about absolute commits, but in order to have a better representation you must spend a lot of time analyzing commits, which at the moment is better used by developing new features and patching bugs. Please do consider that most of the commits at Apache OOo have been about the so called “IP clearance”, i.e. they have replaced “copyleft” code with “permissive” code, and not about developing new features.
Only in the case of the SVG import filter, this has turned into a new and improved feature, but in all the other instances this has not improved the former situation (and in the case of fonts, for instance, has made it worse, as the number of fonts embedded in Apache OOo is now limited by the license and not by the quality and usefulness of the fonts, which – in my opinion as a user – are more important than the license).
Need help with line spacing. Help says it can be done but I have looked and looked and just can’t find where it is in the Writer. please help
Thanks
Links doesn’t work!!! Please fix.