15

Developmental state of LilyPond

The official homepage has a notice that reads "Since no developer currently is listed for commercial development, your best bet is asking on the developer list". And it has been three years since the last stable version, 2.18.2, released in March 2014.

However, there is a beta version, 2.19.54, released January 2017, so I am confused. So, there are no regular developers, but there are still devoted volunteers to write code? Does this mean LP is still being developed?

Furthermore, what if LP becomes abandonware in a few years? I do not worry the same for Sibelius and Finale, since there has been a large user base; and even if they really become obsolete one day, someone will certainly write a converting parser for the many legacy works saved in the Sibelius and Finale formats.

Acceptance of the LilyPond format

I worry about this very much, because this leads us to another question of whether the LP format is or will be acceptable in the music circle. If someday I publish my score (though this is absolutely far-fetched as of now), it takes some courage to save one's original work so dear to oneself in a format people seldom use and cannot open, even if LP was (or is) that good. When people publish, publishers seldom let writers or composers edit source on their own, but call for a format they can handle. The possibility is remote that a rare format like LP's will be approved.

Indeed, there is a library that enables LP to import music XML as LP format (see: importing Music into LilyPond from LP's blog). But not the other way around, i.e. LP producing music XML (see possible output formats in LP manual)—this seems to be impossible.

Moreover, my initial impression is that Music XML files are extremely verbose (perhaps for sake of unambiguity?), and one is unlikely to manually edit them. And it is not clear to me that how can Music XML save my tweaking of slurs and other minute details intact.

  • 4
    No difference between Lilypond and any other open software, which does not fund developers. The format is quite stable, but being a programming language continuously evolves. MusicXML (as all other XML variants) was never intended for manual creation but for program/program communication. Any alternative is better than editing this. – guidot Feb 12 '17 at 12:29
  • Really.... I think I just had the prejudice that commercial software company is more likely to hire developers because there is financial incentives, but you somewhat just convinced me. – Aminopterin Feb 12 '17 at 12:57
  • One way to semi future proof your work is to save as PDF as well as the native format. It's read only but at least you will be able to transcribe into some future format if the worst case scenario happens. – Todd Wilcox Feb 12 '17 at 14:15
  • 1
    @ToddWilcox You can turn PDFs into Music XML using PDFtoMusic. – user1803551 Feb 12 '17 at 20:22
  • 2
    @user1803551 How well does that work? At this year's fosdem I visited a talk by the musescore developers who are doing a project called openscore where they manually transcribe music because they say no music transcription software is accurate enough - yet. One of the goals of the project is actually to provide training sets for music transcriptions software. – Sumyrda Feb 12 '17 at 21:06
20

The chance that "someone will write a commercial parser for Sibelius or Finale," if they are abandoned by their current owners is precisely zero. These file formats are undocumented and covered by intellectual property agreements. Also, the file formats themselves do not provide a full description of what the score looks like on paper - to do that, you would also need to reproduce all of the logic built into the programs themselves (which is also commercially confidential, of course).

On the other hand, Lilypond is open source software, so everything is available at no cost to anyone with the time, patience, and expertise to make use of it.

Lilypond is certainly being actively developed. If you subscribe to the developer list, you will see almost continuous activity.

If you want to preserve scores "for ever", the best option is to save them in as many different formats as possible. The PDF format for computer documents isn't going to die any time soon, and is much more widely used than a specialist niche like music publishing. For music scores, there has been an attempt at creating a standard format, namely MusicXML, though there are some problems with that, both at the conceptual level of "how the musical content is defined" (the standard tends to focus more on the physical layout of the notation on a page than on its semantic meaning), and at the practical level because many music applications that claim to export MusicXML do a pretty poor job of it. The documentation of the first versions of MusicXML was pretty vague as well, which has added to the general level of confusion and misinterpretation.

But even "official" standards like MusicXML don't last for ever. It has already been through three more-or-less-incompatible versions, attempting to fix up the deficiencies in the first release, and a completely "new and improved" next-generation standard is now being developed - though the timescale for actually releasing it isn't yet clear.

Lilypond is in the minority of music applications that don't write MusicXML (and its supporters might argue that there is no point, since why would you want to export data from the best quality music engraving software into something worse?) but it can import it.

Nothing in the computing industry stays fixed for ever. Even in something as commonplace as word processing, who still uses IBM's DisplayWrite, or Apple EasyWriter, or Wordstar, or WordPerfect etc? They were all industry standards in their day.

  • 1
    The question doesn't say "someone will write a commercial parser", but "someone will certainly write converting parser", which is actually reasonably likely. Many undocumented formats have been reverse-engineered, and intellectual property agreements are unlikely to be enforced if the official software is abandoned. See for instance documentliberation.org – IMSoP Feb 13 '17 at 16:18
  • Even in something as commonplace as word processing, who still uses IBM's DisplayWrite, or Apple EasyWriter, or Wordstar, or WordPerfect etc? They were all industry standards in their day. And then there's TeX, which has been with us since the 70s and shows no signs of being replaced by something else but iterations of itself like XeTeX and LuaTeX. – Tobia Tesan Jun 13 at 10:25
20

There are no commercial developers listed for LilyPond right now. That means that you cannot reliably throw money at a particular problem in order to make it go away while relying on the expertise of developers already familiar with LilyPond. Now how reliably can you do that for proprietary software? The problem is that the proprietary software is not available to third-party developers, and developing capacities are rarely available for sale, so you need real muscle money to convince the producer to deal with your problem.

But if you have real muscle money, you are going to find lots of third-party developers willing to work on LilyPond which is available to them legally.

For example, you'll not be able to put any money into Sibelius. Sibelius had been bought out from its original authors (the Finn brothers and their company) by Avid and when Avid got into financial problems, they fired the Sibelius development team in Great Britain, including its founders. The founders tried placing a buyback offer to Sibelius but were turned down. So they are no longer allowed to work on their own software. There is a team in Ukraine paid by Avid to basically keep it running on changing operating systems. But the principal typesetting team has been laid off and could not further work on Sibelius if they wanted to and you offered to pay them, and most of the life support team still allowed to touch the code would not be able to substantially improve Sibelius either, no matter how much money you offered.

I don't see how this puts LilyPond in a worse position than Sibelius, frankly.

As to your editing problems: they are basically inherent in LilyPond being a compiler and working on text. Frescobaldi can do things like transposing passages (most of the time it understands the source code well enough for that) and clicking in the preview will take you to the corresponding source code.

If you want an editing workflow not working with text but a semi-WYSIWYG rendition, try out Denemo (a GUI music editing application using LilyPond as its backend). Of course, not being able to access the canonical source in text form does not have only advantages. But stuff like "delete measure 7 in all voices" will work better there.

  • 3
    I very much agree – open source software can't really become obsolete while there are programming-savvy or pay-willing customers (alas, the latter are seldom found around projects like LilyPond), whereas proprietary software can easily be squelched out of existance by a single company. However that Avid story sounds a bit crazy to me, to you have references for this? – leftaroundabout Feb 12 '17 at 11:45
  • @leftaroundabout Sibelius used to be private company with basically one product, until its founders and owners sold it to Avid in 2006. Avid later decided to rationalize their entire software development activities, fired the complete Sibelius development team in 2012, and outsourced further development to Eastern Europe. The Sibelius team were all hired by Steinberg (though one member of the team decided to go elsewhere) to develop a new notation program starting from scratch. The first version of that has now been released: steinberg.net/en/products/dorico/start.html – user19146 Feb 12 '17 at 12:22
  • 1
    ... The outsourcing to Eastern Europe took a turn for the worse when Russia invaded Ukraine, so Avid then moved their operation for Kiev to Poland, and later from Poland to Canada! All that was happening against a background of reducing their in house staff by about 20% per year, in an effort to save costs. Not to mention that Avid's shares were suspended for a while because of accounting irregularities, prossibly caused by trying to hide some of their financial losses... – user19146 Feb 12 '17 at 12:26
  • That means that you cannot reliably throw money at a particular problem in order to make it go away Lies, if the sum is princely I will do it :) – Tobia Tesan Jun 13 at 10:25
8

You are looking at a page meant for people who wish to get a new feature into Lilypond (or fix a bug) and are willing to pay for it. It is named "Sponsoring" after all.

The page currently lists no developers who are readily available for such work, and instead instructs you to ask around on the mailing list for interested developers.

This sort of sponsoring process is pretty standard procedure for medium/large open source projects. This has no bearing at all with the current development activity.

7

Furthermore, what if LP becomes abandonware in a few years? I do not worry the same as for Sibelius and Finale, since there has been a large user base.

Actually, it's interesting that you say this, since Sibelius actually basically has been abandoned! Much of the staff is now working on Dorico, but Sibelius is now largely abandoned. This is always a risk you take with software, but I think the open-source aspect of LilyPond, complete with the ever-growing userbase, will help prevent LP from being abandoned anytime in the near future.

And by "commercial development," I understand that there are no developers currently being paid regularly to improve the program. Being open source, LP is being actively improved (these are the updates that come out pretty regularly), but on a volunteer basis by the community.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.