Update on a Blogosphere Quickie
Original Mention, Blogger’s Reply
We have yet another reply to a mention here on the ACOP blog, this time from the user who had a funky file that only opened in Notepad, and which didn’t open in AbiWord. I would agree that notepad is a relatively mature product, but this is made easier by the fact that it is very simple. (The whole concept of import/export is beyond it, as it is simply a plain-text buffer that only complains a little if you try to open a binary file in it, and it barely handles encodings). I don’t think it’s undergone that much revision - I mean, the Ctrl-S shortcut for save was added in Windows 2000! Other text editors for Windows, such as metapad, are much more featureful and useful and just as “mature” - a plain text editor is a whole lot easier to write than a full word processor.
AbiWord import/export support is a constant work in progress, and we have a Summer of Code student funded by Google working right now to make a lot of improvements - many of these have already shown up in 2.4.5, which is available now.
I have a minor factual quibble with the reply - AbiWord has actually been around since around 1998 (that’s the earlier copyrights on the code - my guess for where the 2002 number came from is that version 1.0 (which followed a series of 0.7 and 0.9 releases which were highly useable), was released then).
As one might guess, I would also urge that the “numerous missing features” as well as the original document that caused the problem (or a similar one exhibiting that or any other problem) be filed in the AbiWord Bugzilla. If those missing features were obvious to everyone, someone would have implemented them by now most likely 
since lots of people use AbiWord in their day to day work and play. One thing that I’ve found in my work in personal end-user support is that everyone has their own idea of how everybody else does things and what the “only logical” way to do something is. This is not a criticism, just an observation that anecdotal, subjective opinions about usability matters are less valuable for developers, yet much more important for those responsible for training. (You’d never believe how many different ways there are to dial up to the internet and check your email on Windows, for instance) With reference to the RTF import bug, if the document truly was valid in some reasonable editor, we should be able to at least make headway in importing it. If it’s not valid, under no circumstances should we crash or hang, and trying to display some data at all is better than giving up. (The LibWPD WordPerfect import library recently went through a strenuous QA review by one of our QA folks, Sum1, and so it is now a shining example of this ideal, opening more corrupted WordPerfect documents than even WordPerfect itself.) I think a general judgement about the “maturity” of AbiWord is misguided, as the design from day 1 was very well thought out and mature, and the implementation has been “mature” enough to use for nearly all tasks since late in the 2.0 series or 2.2, IMHO, though I know of people who thought it was mature long before then. One import error does not a word processor define - that’s what bug reports are for.
Regarding the Mac OS X rendering, yes, we know this is a work in progress. We have only one active Mac OS X developer, who (like all of the core developers) volunteers his time (unlike the vast majority of OO.o developers, who work on the project through their employment with Sun, Novell, etc). Unfortunately he is very busy at the moment. Contributions are always welcome, and if you have any experience in Mac OS X development and would like to help out, please drop us an email or drop by the IRC channel, and we’ll see what we can do to get you going. (As a work around, you may use the latest 2.2 release - while it is much less featureful and has numerous bugs that have been resolved since, the text rendering quality is better. This is a stop-gap measure at best, though.)
Thanks for taking the time to elaborate, Scott! Feedback, especially in the form of bug reports or code contributions, is greatly appreciated! (BTW: Any impression of a “tone of disbelief” was most likely mis-interpreted terseness due to the number of blog entries I look through to make these posts, combined with a desire to see a test case document in Bugzilla so that the problem could be resolved. Sorry about that!)
Update on Docvert!
I love it when those I blog about show up and say hello!
Via email (thank you for your persistence - my email has been added to the About the Editors page…):
(hope one of these emails gets through — couldn’t find a contact link on cleardefinition.com)
> In the field of “things to look into”, this blog post pointed out something called Docvert, which I hadn’t run across previously. It seems like a pretty commonly-requested web document conversion service, this time specializing in producing ODT files, which AbiWord certainly can do. The blog post mentions AbiWord support, and the download page requires either AbiWord or OO.o, as well as the PHP scripts available. Would be interesting to see how it uses AbiWord - I wonder if it uses the spiffy (apparently, never used it myself
) AbiCommand plugin, or simple command line conversion… It would be great if people who do things like this would let us know and give us their feedback, it’s the open source (or free software, no flamewars please) way. Check it at http://digitizationblog.interoperating.info/?p=309 and http://holloway.co.nz/docvert/index.html
Hi, I’m the guy who runs the Docvert project, and yeah I guess emailing the people whose work I build upon is a good idea
The way Docvert builds upon Abiword is by getting it to do command-line conversions from MSWord to OpenDocument. It doesn’t dynamically build up the document using the AbiCommand plugin — instead it’s more about converting A to B.
After this conversion Docvert takes over and converts ODF to DocBook 5.0, and then HTML/XML*.
It can use OpenOffice.org for conversions, and from my perspective it’s about using Abiword and OpenOffice.org as mature conversion libraries. In this way users choose which conversion library is best. I’ve found that OpenOffice.org is quite good but it’s hard to keep going, what with the requirements for XVFB and all.
Anyway, thanks for telling people about the software. Cheers dude
* Going directly from OpenDocument to HTML would have been a pain in the arse… DocBook’s hierarchy of tags is easier to write unittests against (Eg, assert that all sections have titles) and it was much easier to deal with than a flat list of paragraphs and headings, a la ODF and HTML.
.Matthew Cruickshank
http://docvert.org
It might interest Matthew to know the recent improvements (nay, overhaul) of the Docbook import/export plugin courtesy of our Summer of Code student Sum1. It may also interest him to know that while AbiCommand does allow construction of docs, it can also be used for simple conversions (from Whatever to ODF, DocBook, or right to clean XHTML/CSS!) without spawning a new process each time. Google around for AbiCommand information - there should be some on Martin’s blog as well as in a nice “+site:abisource.com” Google search. I’ve never personally used AbiCommand, but I have heard of lots of folks who really like it. A bunch of patches just went in and will show up in 2.4.6 as well as the 2.5.0 development release.
Read and reply, ants!