the state of tablet software in linux

Posted on by Luis Villa

I bought a Lenovo X41 tablet two weeks ago. I got it for a number of reasons- ease of note-taking, improved interaction with professors, just needing a new machine anyway, etc. I knew going in that the state of linux tablet software was sort of poor, but after more investigation I’m surprised at how poor, and perhaps more troubling, how fragmented and nearly stagnant it is.

There are a variety of notepad-type tools. Jarnal came with the tablet, and is pretty damn good, except for the handwriting recognition, which is passable with training, but not great. Unfortunately, it appears that there is no active development, so we get what we get. Gournal is around, and looks pretty solid, though I have not actually used it yet. Xournal looks to be both cool and actively developed, at least for now, though it is not as sophisticated as Jarnal (no handwriting recognition, primarily; I have great faith that gocollab will make the collaboration part trivial soon ;). Maybe Xournal will be my first review for Quim’s new project– I’ve been using it every day for two weeks now and am pretty happy with it. None seem as polished as their Windows counterparts at first glance, but they are decent, and Xournal may be picking up steam. Sadly, Xournal appears not to be part of edgy, so I’ll likely have to break my no-compiling rule for another app (currently only gimmie is cool enough for that.) (If tomboy allowed scribbling into notes as well as typing, I’d wet my pants with joy, by the way.)

Unfortunately, for notepad-like tools to become really, really useful, they need handwriting recognition. More unfortunately, the more interesting-looking handwriting recognition/stroke-based text entry projects all appear to be dead in the water. GPE’s rosetta hasn’t had a significant commit in 16 months; mallum’s cool-looking xstroke replacement hasn’t had one in 11 months; and herzi’s HRE work got very little interest at fundable. This is a shame- I wish I had some clue about the ‘why’ to this problem- I’m guessing some combination of low demand and high skill barrier, but don’t really know. (Is there some project out there I don’t know about?)

There is at least some push in the keyboard area- obviously GOK is powerful, omnipresent, and under development. Frankly, though, I was surprised at how difficult it was to get going, and it apparently disabled my pen at one point! I’m intrigued by Ubuntu’s ‘OnBoard‘ project, and it looks nice and ‘Just Works’, though I’m curious why it isn’t just a GOK mode- just too radically different in aims? The laptop shipped with xvkbd, which works, once you figure out the trick :)

It doesn’t appear possible with any of these tools (either keyboards or stroke-recognizers), at least that I’ve seen, to use them in a maemo-like mode, where tapping on an input field brings up a text entry tool. Maemo’s behavior in this respect is so obvious and intuitive that I’m shocked I haven’t seen it elsewhere. It should certainly be in every linux tablet setup.

Anyway, I do think that tablets are going to become more popular, especially once text entry gets solved. Hopefully I’ll be along for the ride :)

[Coolest non-obvious feature I’ve seen on the windows side, by the way? Onenote will record audio and tie that to your notes, apparently, so that you can ask ‘what was being said when I wrote this?’ and get back that audio section, to use to polish/expand on your notes later. Very nice hack.]

[PS. I really hate WP’s text editor’s handling of white space. Does anyone know if there is a fix for it?]

Published by Luis Villa

29 thoughts on “the state of tablet software in linux”

  1. awed by the pressure-sensitive effects of the GIMP and the tablet stylus. Later in the afternoon we installed Xournal and that provided us with a second round of excitement. The state of tablet support sdaly seems to not have changed a bit sinceLuis summarized ittwo years ago. Namely, hand-writing recognition is still a completely missing piece. All in all, the booth was a great success. Compared to last year, I think we got, like, five times the traffic at the Fedora booth, if not more. Was a pleasure to meet old and new friend

  2. Luis; cool post, some comments.

    o Arg, matchbox-stroke basically works as a simple grid based recogniser, it just needs a decent config file written for it, ideally by porting an old xstroke config to matchbox-strokes xml format. This being a tedious though relatively un technical task, myself not being a massive fan of stroke recognisers and just generally always being over busy it never got done :( I assumed some brave soul would one day attempt this and a patch would appear in my mail. Hasn’t as yet.

    o One of the other reasons why there arn’t many o/s stroke recognisers is I believe there are a whole load of nasty patents covering stroke recognition.

    o All the Maemo software input method foo is closed source. The stroke recogniser engine, I believe, is actually a 3rd party commercial engine.

    o I think the GPE folks have a GTK hack to signal the opening of an input method on key input focus ( like maemo ) but like the Maemo mechanism this isn’t very generic ( iirc ). I hope the GMAE effort can work towards standardising mechansims like this ( similar to what f.d.o has done with desktops ) so the same input method can work across a GNOME tablet, a maemo internet tablet and a GPE handheld. There all X+GTK based after all.

  3. * Are there any instructions posted as to how the xstroke->matchbox-strokes porting is done, mallum? I’m in law school and was in QA; I specialize in tedious and relatively untechnical.

    * I figured patents were an issue (pursuant to Palm’s patent troubles); didn’t want to speculate completely wildly on that.

    * All the Maemo software input method foo is closed source. The stroke recogniser engine, I believe, is actually a 3rd party commercial engine. Ah. So I should have said 770. Ah well. (They need to fire their 3rd-party commercial engine; I find their stroke recognizer to be the worst part of the 770’s software stack. I just like how it is activated.)

    * It would be great if that were standardized. Obviously there isn’t enough demand right now for anyone to pay for tablet development, but if tablet could piggyback off GMAE, that would be great.

  5. Admittedly I have not looked at basket in a year, but frankly, it struck me as the wrong direction at the time- they should have called it ‘kitchen sink’. There was nothing useful in it that couldn’t be done in tomboy by giving tomboy a slightly richer notion of objects within it. (Mind you, tomboy needs some of these things- like outlining- desperately, so at this time, basket might be more useful for many people than tomboy. But I don’t see anything useful there that tomboy can’t get while remaining simple, and I see lots of unnecessary, useless complexity that thankfully tomboy will never have.)

  6. Evernote (http://www.evernote.com) actually has a SDK of riteForm available for Linux. AFAIU it could be used to write an HWR interface. It uses the riteScript HWR engine.

    The missing HWR on Linux (and good integration of GOK) is the reason I still rather use Windows XP Tablet Edition on my UMPC, though Ubuntu runs on it fairly well (didn’t test all as it doesn’t make sense for me to use a Tablet PC, or UMPC, without HWR).

  8. @luis

    – what needs doing is;
    – Grab the original xstroke config file ( uploaded a copy to http://o-hand.com/~mallum/alphabet )
    – The matchbox stroke is here; http://svn.o-hand.com/repos/matchbox/trunk/matchbox-stroke/data/strokes.xml
    – Hopefully the conversion is quite self explanatory, basically the contents of xstrokes grids()’s
    need to become tag ‘sloppy’ attributes.
    – Drop me a mail if you have any problems.

    As for tablets piggy backing on GMAE, Im really hoping this kind of thing can happen. Tablet XP for example must use some technology from pocketpc.

  9. I’m the developer of onboard. I decided to write a separate app instead of extending GOK mainly because the gok devs wouldn’t allow me to make the changes to let it work with no configuration and I felt this was important.

    thanks for the mention :)

  10. My take at notes taking, was horizon (http://pippin.gimp.org/horizon/) mostly done as an investigation into capabilities needed for a buffer implementation for GEGL. It is a sketching application that provides an unbounded canvas that is pan and zoomable.

    The thing also provides the ability to import/export pngs/jpgs as well as creating hyperlinks within the canvas (these shows up as thumbnails).

    It was originally written for the Nokia 770 which was not high powered enough to do nifty animations when following hyperlinks. A desktop like PC on the other hand allows such software accelerated graphic effects to take place though. Horizon as a project is currently dormant, awaiting a more performant incarnation of the 770 or something similar, and a resurrection of the GEGL branch (which only exists in some tarballs I keep around from march of this year).

  12. Mauricio, Oyvind: thanks for the pointers, I will take a look at them when I have a chance.

    Chris: shame about that; it does seem like a reasonable request on your part. Is there a link to that discussion somewhere?

  19. […] a tree model for notes. Basically, what I want is Tomboy with hand drawn pictures support. Update: Luis Villa wrote about tablet software for Linux, and one of the subjects was this same thing. One program he mentioned that is really interesting […]

  22. […] xournal is available through Koji, which is awesome. Gimmie is packaged too- more awesomeness. (Overall, sadly, little has changed on the tablet software front since last year.) […]

  23. […] Luis Villa’s Blog / the state of tablet software in linux Luis Villa’s Blog Ramblings on law school in New York, free software, and the spaces in between. Skip to content Home the state of tablet software in linux I bought a Lenovo X41 tablet two weeks ago. I got it for a number of reasons- ease of note-ta […]

Comments are closed.