The jBrowse FAQ

Over the years, a fair few questions have been asked about jBrowse. jBrowse is no longer under active development, but it is under active use, so I am collecting a few common questions here.

Why no Opera/Mozilla/Firefox version?

This is by far the most commonly asked question about jBrowse. The answer is: it's too difficult. I agree that having jBrowse only in IE is a huge pain. I use Opera myself. But implementing jBrowse in Firefox (which is the second most widely used browser and the best understood one for plugins) is unbelievably hard.

Challenge 1: there's no ruby tag in Firefox (as far as I know, at the time of writing). So there goes half of jBrowse's functionality right there.

Challenge 2: XUL. Firefox uses a very very complex system called XUL. How complex? Well, jBrowse consists of some source code. That's it; a bit of C++ code that implements a documented interface defined by Microsoft. Making source code is easy, especially when it basically gets given the root node of an HTML document and then simply has to manipulate that document and hand it back. That's all jBrowse does. The simplest possible XUL application is probably about as complex as the WHOLE of jBrowse, and contains more separate files and requires more knowledge to create.

I don't know if that says a bad thing about Firefox or not, but life is simply too short to create a version of jBrowse for it; I'd rather wait and see if IE7 is good enough to use, which by now I think it is.

How can I update the dictionary?

jBrowse comes with a rather old version of the EDICT dictionary file, compiled by Jim Breen and his minions. For speed (not that it makes that much difference in fact) jBrowse's copy of the file is in UTF-16, as opposed to the UTF-8 that the original was in. To update the file, therefore:

• Go to the EDICT site and get a new copy of EDICT. Don't get the gigantic XML version. jBrowse can't parse it and it's a painfully inefficient format anyway.
• Convert your file to UTF-16, using the handy tool of your choice, such as Microsoft Notepad or vim.
• Save it over the existing dictionary file, which will be called edict_ucs2.txt
• Restart jBrowse and make it re-load its dictionary. The new words will be there.

It's also possible to just add your own words to the dictionary. It's also possible to update the Kanjidic file, but that file doesn't change much.

How does it work?

In terms of how it works as a browser plugin, I can't really be bothered to describe it; it's not pretty and it's not easy but it's not very interesting either (that just about sums up the Internet Explorer interface, in fact).

In terms of how it analyzes the text in the browser, that's described here.

Why does it keep showing rare words when there's an obvious non-rare alternative?

EDICT doesn't tell you much about whether a word is rare or common. EDICT is quite a piece of work and it's free so I think we should all be very grateful for the information it does contain -- however, it is really really annoying that sometimes jBrowse will just use a very rare word that comes first in EDICT rather than the common word we all know is correct. This can mess up a whole sentence sometimes.

There are things you can do about this though -- in particular, you can remove a lines from the dictionary file (see above) if it is causing 'false positives' in jBrowse.

Does anyone actually send any money?

Most don't, and that's fine, but some do. It's my intention to create a Hall of Fame of those who have registered jBrowse, at some time Real Soon Now.