The LilyPond Report #9

This week’s issue discusses open-source hardware, accordion support, intermediate files, LSR downtimes and bug statistics. I’d also like to announce that next week’s issue will be a very special “LilyPond Report”, entirely dedicated to... oh, you’ll see. As always, you can post your comments at the bottom of the page, or even register and contribute to the LilyPond Report’s next issues.

 This Week’s Desultory Editorial

Greetings,
well well, this is already the tenth issue of our LilyPond Report (since the first issue was numbered as #0), and so far I think the experiment has been pretty successful... and the success is actually yours much more than mine: without all the interesting conversations, ideas, contributions that are constantly to be found among the LilyPond community, none of this would be possible.

Of course, keeping up for ten weeks or for fifty-two weeks is a completely different story, and I have no idea whether the LilyPond Report will still be alive in 2009 or not; honestly, this is the kind of questions I’m trying to avoid at all cost right now.

So-called “Newsletters” mainly exist in large projects such as GNU/Linux distributions, and most of the time these are published on a monthly basis, or are kept sparse (sometimes these are nothing but a collection of links). This is why, although Graham Percival once referred to the LilyPond Report as a “newsletter”, I never intended to actually run a “proper” newsletter, and I hope this way it will be more fun, more flexible, and more likely to last.

To see what kind of things I have in mind, please come back next week and have a look at the official #10 issue: thanks to some distinguished LilyPonders, a very special and interesting Report is indeed being prepared. Stay tuned!

 News from the Free World

Software can be free, we already know it: using a GNU/Linux distribution you can have a whole Operating System entirely free (as in “free speech”). Drivers are a bit more problematic: for instance, your 3D accelerator chip or your WiFi card may require you to use non-free drivers or firmwares. However, can the hardware itself become free?

For the past decade, countless initiatives have been aiming to make it happen. The OpenCollector project is trying to list them, though their diversity makes it unbelievably difficult. Interestingly, this website refers to the Hippie movement more than once, and mentions that the tradition of Homebrew Computers dates back to the 70s.

Open-source hardware includes amateur and professional projects, from plain do it yourself to huge communities or corporations. Exactly like free software, these projects have in common an ability to exist without any guaranteed income, a cooperative development that relies entirely on the Internet and a constant commitment to making their specifications available to the public.

Open-source MP3 players are generally proposed as DIY kits: let’s mention, for instance, Frank’s, Butterfly and other kits.

Open-source game consoles are a bit more involved. You may know about the Hydra kit, that is absolutely not free; I’d like to mention the French DOGS project, that looks quite interesting, and, on a much more professional-looking level, the GP2X. We can also expect more projects to appear in the future.

Open-source graphic cards are very promising. The OpenGraphic project, started by Timothy Miller, is showing us a promising path; their first chip is now ready and available for pre-order. However, even major companies seem to make their specifications available, so things are moving really fast here.

Open-source processing cores, similarly, are now fully functional. The OpenCores project has been started to achieve “the ideal of freely available, freely usable and re-usable open source hardware”. As noticed in a Free Software Magazine interesting paper, “Open hardware goes mainstream”: even the Sun Microsystems company has made their Sparc processor publicly available, under the name of OpenSPARC.

The UltraSPARC T1 chip

OK, I could stop here. BUT... this is precisely where things are getting more interesting. What if... anything could be free (as in “freedom”)?

We already talked about Free Beer. There are also
 Free Robots,
 Free Prosthetics,
 Free Particle Accelerators,
 Free Bikes and Bicycles,

The C,mm,n prototype

 no less than three different Free Car projects: the Open Source Green Vehicle, the C,mm,n project, and the OScar — all three have impressive designs!

 And tomorrow... Free Space exploration?

While the world we know is often about short-term benefits, there are people working everywhere to invent something else, something new, something Free.

In two weeks, we’ll have a look at a project Andrew Wagner told us about some days ago: the Do-It-Yourself keyboard.

 The LilyPond Feature of the Week

Last week we had a look at LilyPond’s support for Bagpipe; this week I’d like to mention support for Accordion.

As Graham noted in a recent mail, both “the bagpipe and accordion stuff [...] have no special code-features; they’re all done with”extra“.ly files”. (You can have a look at these files in your LilyPond installation directory; however most of the time, you’ll just want to \include them in your score.)

However, this is only partially true, since the LilyPond “Feta” font itself includes some special glyphs for the Accordion. Accordion support is not well documented yet; however you can have a look at some LSR snippets to see some possibilities:

 Diatonic accordion (this snippet includes a macro for jEdit)

 Discant symbols

 Hopefully more to come...

A few months ago, David Kastrup explained in a long mail how support for accordion could be improved. His request has been added to our tracker, and is now waiting for someone to handle it... If any skilled programmer/accordionist ever reads this, well, you know what to do!

 The Idea of the week

This week’s idea has been around for quite a while.

As you may remember, LilyPond originally produces Level 3 PostScript files, with a .ps extension. These files are afterwards turned into PDF files, that most people are much more familiar with.

So, here’s a simple idea Laura Conrad once mentioned, before Graham summed it up:

I agree with Laura: we should treat the .ps files as temporary and delete them.

[...] I think that deleting the .ps files is a good default option. Most users don’t want ps, and many users who investigate the ps files won’t know how to deal with them properly. Anybody who really wants a ps file can invoke with —ps.

Indeed: seeing “exotic” files appear on your computer is always confusing for new users, particularly for Windows users who not only get an unwanted .ps file, but also a .log file as well.

Granted, there’s already an option that does the trick. This require to launch LilyPond from the command-line:

Launching LilyPond from the command line

notice the —ddelete-intermediate-files option

Mmm... Not very handy, is it?

 The LSR-downtime of the Week

Some of you have noticed that the LilyPond Snippet Repository was going through a rough time.

As you may know, the LSR is developed by Sebastiano Vigna and gracefully hosted at the University of Milano, Italy.

 On May the 1st, a broken router affected the LSR server. It was corrected in less than 12 hours.

 On May 6th an update in the Java code behind the LSR made it unavailable again; it was fixed a couple hours later.

 On May 11th, the LSR database got flooded by a too important amount of connections. It was fixed within 6 hours after reported.

This tells us two things: the first one is that we need the LSR more than ever: every time anything goes wrong, LilyPonders notice and report it. Secondly, this proves how dedicated and reliable Sebastiano is: well done Seba, and keep up with the good work!

 The Bug(s) of the (80 past) Week(s)

Several days ago, Graham Percival has stated that “we have a fair number of new people fixing some bugs”. I don’t know exactly what (or who) he meant, but I have to say that these past few weeks, I have indeed felt that we were achieving some kind of balance between the bugs appearing bugs being fixed — was this feeling justified? I had to investigate this.

Now for some statistics. Our Google bug tracker was opened on August 20th 2006, but for some time it only hosted fake (but amusing) test reports. It wasn’t really used until 18 months ago, where it was interfaced with our bug mailing list.

Here is a histogram showing:
 in red, the total number of mails sent to this mailing list
 in blue, the number of mails that were automatically sent from the tracker
 in yellow, the ratio between these two curves.

Statistics of the bug-lilypond list

From October 2006 to May 2008

Hmm. Does it prove anything? I’m not sure. Mails are sent by the tracker every time an Issue is created, fixed, or simply commented, so these types of data lack reliability. However, we can see three distinct bumps in January 2007, in September 2007 and in January 2008, that affect the tracker activity as much as the list.

There’s something else. If you look closer at the histogram, you can see that, while the blue and red curve have been slowing down for the last four months, the yellow curve is starting to go up, as of April 2008. In other words, for the very first time, there is more activity on the tracker (i.e. more bugs discussed and/or fixed) than on the mailing list!

For example: since May the 1st, 4 new issues have been opened, while almost 30 issues have been dealt with on the tracker (most of them have been fixed; on May 1st only, Han-Wen handled about 20 issues on his own! — mostly ties and slurs-related, if I remember correctly).

So, is LilyPond in good shape? I do think so — even if we lack convenient tools to tell exactly how good.

And this concludes the ninth issue of The LilyPond Report.

Cheers,
Valentin Villenave