TODO

Intro

it contains a list of things related to the web interface, and a list for the rest.

Web-related issues

1. Conversion of input (today only c1, d1):
   1. Set up test environment (Boerre)
   2. Test for {MS/Linux}-Unicode and for Mac input/output (output may be easier to control, there is a "print result as html" section in the cgi-bin script.
2. Make an interface for chosing tag language
3. Generating: Make a scheme for ticking off grammatical alternatives, and a cgi-bin-script for generating forms according to the choices made.

Technical things

1. Bugzilla, the bug database (Boerre)
2. The scripts must be documented
3. Conversion scripts for different code pages
   1. There must be conversion scripts for all relevant code pages
   2. These scripts must be in gt/script, and under cvs
   3. They must be made for perl, and (eventually) for xfst
   4. They must be documented, and tested
4. Conversion scripts for html and pdf to 7bit
   1. There must be one html script, containing all types of tags
   2. Again, this script must be in the script directory, under cvs
   3. Similarily for pdf (if possible, otherwise several pdf scripts)
5. Conversion scripts from 7bit to html (cgi-bin use)
   • These could perhaps be done as part of the cgi-bin files.
6. Documentation for the scripts !!!
7. There should be parsers for tag output (not abbreviations, tags in Sámi, etc.)

Bugs

1. The sma and smj Makefiles generate two binary tok.fst files instead of one. One of the two is placed in the bin directory, as it should (and everything thus works fine), but the (identical) copy is stored in the src directory itself (and it gets a notice from the cvs). Nothing in the makefile hints at such a behaviour, so anyone interested in ghosts should have a look at this issue.