These are things that still need to be taken care of. Items that are checked (i.e. preceded by `v') have already been finished. Starred items await completion. * Remove all obsolete code * Fix all code marked FIXTHIS v Add $success_flag to process_command (currently, if in trial mode, process_command signals the calling function a failure by returning undef. This complicates all functions that call process_command since they have to deal with this possibility. Using a flag instead is much cleaner) v Need to think about how to handle add_to_latex_body for arguments that are processed by process_command. Since this comes up often, maybe a process_argument routine would be useful that acts as a wrapper for process_command v Make ALT text for footnotes more meaningful v In general, make ALT text contain LaTeX code for equations v Finish implementation of bgroup and egroup (check for proper closing of braces and environments) v Finish implementation of begin and end (check for figure and table environments and extract captions and labels if necessary) v Reimplement tables * Add LaTeX2e tables (m, b, >, <, and ! options) v Finish up the ISO-8859-1 table (e.g. add times, divide, etc.) v Make accents use two characters if necessary (e.g. ¯A if Ā is not available) v Add option so that some characters will be processed in LaTeX if necessary (i.e. for crazy accent-letter combinations not available in the font, create an image) v Implement those math entities that are available in ISO-8859-1 (\pm, \times, \div, \ast, \cdot, \ll, \gg, \neg, \mu, \backslash) v Add log-like functions and \bmod and \pmod * Fix limit-like functions to process subscripts correctly v Commands (not just environments) to be processed in LaTeX * Add remaining AMSTeX math entities * Add lots more comments and arrange the subroutines in an order that is pedagogically valuable. Write comments in a format that makes easy extraction of documentation possible. v Rewrite the style files * Need to think about equation numbers (heqn.sty) v Check if list of figures, list of tables, bibliographies, glossaries and indexes work right * Check if footnotes work right v Implement \sl and \sf via classes v Install-test should check for perl version 5 and should strip the documentation from the code v Add COMPACT attribute to DL lists v Change process_text to split only if sc or mi are active v Implement comment, htmlonly, and latexonly environments v Implement do_cmd_comment * Make regular expressions robust against empty brackets v Fix process_ref v Add do_cmd_providecommand * Implement rm, sf, tt, etc. environments v Update list environment v Generalize stacks according to Michael's suggestion * Make icon ALT texts language dependent v Use get_image_size to get icon sizes v Make htmlimage print arguments to log file and retrieve the image parameters in process_log_file v Change labels to approximate LaTeX's behaviour: if refered to be \ref, jump not to the label but to the last section/table/figure/etc. but if refered to by \pageref jump to the label itself. This means that certain environments and commands insert named anchors even if they do not contain labels and it means that \refs do not point to the corresponding labels but to whatever it is that is being referenced to. v mbox needs to unset $math_mode v Process counters as regular text and/or implement them in latex2html completely v Redo the do_cmd_today subroutine to be more language friendly v Reduce the number of backticks and system calls v Implement "math shortcuts" $$, \(, and \[ v Apply revd patches * Apply all patches from 96.1 to 96.2 when it comes out. v Fix image caching. Don't use links, instead try to remember mappings. * Implement \ensuremath * Retire %ref_files v Move image generation code into separate script to increase portability. v Obtain page numbers from the log file instead of relying on a counter which might get confused if an image extends over more than one page. * Finish implementation of index (i.e., add makeidx package) v Read only one line at a time