Completed Tasks

See also:
  1. 2000-06-09 Bug can't set targetServer in a subdirectory.
    Need to match and remove a prefix, let's say maybe SITE:docOffset. The parameter to set this would then be offset=path. (2000-06-07)
  2. 2000-06-07 Hole Do a _subsite.xcf file for standalone Woad
    i.e. an Apache's document root. This may be simpler than the PIA version, but will still involve bug fixes to make the top-level /Woad directory work for tagsets and so on. Of course, could put src-*.ts in PIA/lib for now. (2000-05-30)
  3. 2000-06-07 Hole (dps) wrappers, #element, #type in tagsets
    Need to define wrapper element and default bindings for #element, #comment, etc. Default element is handled differently from other node types, which have to be rewrapped as elements by the parser. GenericHandler has to operate differently for non-elements. (2000-05-30)
    -- (element="", documentWrapper, element="#type" 2000-06-05)
  4. 2000-06-12 Hole (site) WOAD directory listings.
    In order to make WOAD-based directory listings we need to map - into something, e.g. IndexDocument. A web page would be cleaner, but it might be simpler to cons up a DOM tree and remap the tagset. The default mechanism does work, FWIW, but does not handle "pia:". (2000-06-09)
    -- The IndexDocumentPath hack works, too, mapping the listing into a Resource, but unfortunately the Resource is processed in the wrong context (i.e. the one it comes out of). The Listing has to hold onto the Resource and reprocess it in the right context.
  5. 2000-06-14 Hole (pia, site) error handler for missing documents

    One way to do this is by passing the remaining pathinfo as a query to the listing document in the last valid directory in the path. Note that this needs work in both pia and site.

    (2000-06-13)
  6. 2000-06-16 Bug rename the path parameters per about.html
    (2000-06-14)
  7. 2000-06-16 Bug target mapping screwed up after changeover
    Neither sources nor notes are properly getting a server link in normal mode; everything but the top-level dir. is ok in source-server mode. Notes don't map at all. (2000-06-16)
  8. 2000-06-22 Bug fix decorate-path
    Should start with a base URL and go down, rather than using ..: that way it will work even if there's a tail. (2000-06-21)
  9. 2000-06-23 Hole quick start must explain various setups
    In particular, how to set WOAD up for either personal or shared annotations, sources. Need to make sure the woad command has the right defaults. Must be able to tell it where the real (annotation) root is [should be the filename]. (2000-06-22)
  10. 2000-06-26 Hole site indexing; links TO pages
    There are two ways to do it: incrementally while browsing the URL tree, and offline using a crawler (similar to woad-index.pl). It's going to be slow unless we use a namespace. (Maybe SITE:siteMap.) Note that after site indexing you can have a "directory" listing at each directory-like URL. (2000-06-21)
  11. 2000-06-28 Hole note listing, editing and creation
    "label" = filename base. When creating a new note, suffix a number to the chosen label. Combine source and URL listings if possible. (2000-06-22)
  12. 2000-06-28 Hole annotation tagset.
    We may actually need several: url-space notes / source notes / indices. (2000-06-07)
    -- (2000-6-21) in progress
  13. 2000-06-29 Impr HEADER note with content appearing at top of page. (2000-06-28)
  14. 2000-07-21 Impr Refactor source viewing
    .li extension -> src-file.ts, ancestral to src-html.ts and src-xhtml.ts. Maybe rename *-listing.xh to dir-*.xh (2000-07-20)
  15. 2000-08-10 Note Watch out for empty files in annotation dir's
    If somebody has done touch ~/.pia/.words/foo and you later ask for /.words/foo you get a ``document contains no data'' error, which is correct but confusing. (2000-08-10)
  16. 2000-08-22 Impr sources -> source on command line
    for consistancy.
  17. 2000-09-25 Hole chronological, xref index for notes.
    in other words, the indexer has to be run over the notes directory. (2000-09-22)
  18. 2000-09-25 Bug woad-index.pl must entity-encode attrs
    Some titles contain quotes, for example. (2000-09-25)
  19. Impr allow arbitrary source=target mapping
    either a .wi file, or a tagset included in woad-xhtml. Needed for stuff like /cgi-bin. Cannot be done in .xcf file because that results in undesirable aliasing. Could be done with an included tagset. (2000-06-23)
    -- can be done with aliases=./..,/..., and offsets=/...,/..., on command line.
  20. 2000-09-29 Hole cross-references for words.
    See hook: TextParser.lookupXref(...). Look it up in a namespace (which could, e.g., be tied to a database). Linking can be done in listing parser if there's a suitable namespace in context. (2000-09-22)
  21. 2000-10-02 Bug problems with word xref indices
    We may want a different expansion for Def and Ref in xrefs. In any case, there's no expansion for Ref at all. Def needs to show, and link to, context. The word link is just plain wrong. (2000-09-29)
  22. 2000-10-06 Goal useable release by 9/30.

    -- internal release for TouchVerse 10/2. Public release scheduled for 10/6.
  23. 2000-10-06 Hole linking of identifiers to annotation, docs.
    This requires two things: proper tagging of identifiers in context, and figuring out where in the tree to link them. Generally, each identifier gets its own page, Wiki style. In some cases the context requires its own page, i.e. for Java identifiers or XHTML tags. (2000-06-07)
    -- (2000-06-28: annotation listing done. Ready for indexing and linking.) The simple thing was to give each identifier its own cross-reference page with an easily-computable URL derived from the identifier itself (lxr does the same).
  24. 2000-10-12 Bug AllNotesByTime needs to be appendable
    so that new notes can be appended as they are created. This means that it needs to be oldest-first instead of newest-first, with a link to the bottom. (2000-10-11)
    -- links: 2000-10-11; woad-index.pl: ; create/edit:
  25. 2000-10-12 Bug link from a note to the page it annotates.
    Noticeable especially in .sources, where there's an extraneous .notes interpolated into the URL. Obviously another mapping function is needed. Back button is fine, but doesn't work if you got there from AllNotesByTime. Could do this with <annotates> tag. (2000-09-25)
  26. 2000-10-13 Bug word list uses too much memory
    The correct solution is to use a simple set (i.e. a Namespace subclass that maps words to themselves using a Table) and compute the links in <id> or else in the parser (simple for now, but basically wrong). (2000-10-07)
    -- the actual solution uses the xprefix attribute and a map that takes, e.g., word into -W-/. The trailing slash tells the parser to append the word, lowercased.
  27. 2000-10-18 Bug word xrefs still use more memory than necessary
    The correct solution is to use a simple Table specified by a new namespace-like tag. TextParser wants a little cleverness to extract the table if it exists. (2000-10-13)
  28. 2000-10-19 Bug TextParser needs to inherit from CodeParser
    currently xrefs is only fixed in TextParser. Actually, they should both inherit from ShallowParser. (2000-10-19)
  29. 2000-10-27 Impr shallow parsing
    REX: XML Shallow Parsing with Regular Expressions provides the regexp needed to mark up an XML file as a listing. OTOH it's within epsilon of what we're already doing in AbstractParser. (2000-08-11)
    -- Current implementation of TextParser is different from either, but it seems to work pretty well.
  30. 2000-10-30 ~ Impr xref C code using LXR
    Why re-invent the wheel? LXR looks great, and it's in PERL. GPL. Doesn't grok Makefile, html or perl, though. (2000-08-07)
    -- not necessary because shallow parsing and indexing are working OK.
  31. ~ Impr Full-text source indexing
    Glimpse is what LXR uses. Unfortunately it's free only to nonprofits. An earlier version (4.0) was free; may be available from the CVS tree. See FreeWAIS-sf for a possible alternative. The original FreeWAIS doesn't seem to have been worked on since 1996. List of full-text indexers in DMOZ. SWISH-E also possible. This article has useful info about doing it in PERL. Turns out that Swish and Glimpse are what Harvest calls "brokers" and BOTH are included in the Harvest distro. But the Glimpse in the current Harvest distro is still non-free. (2000-08-07)
    -- not necessary because shallow parsing and indexing are working OK.
  32. 2000-11-29 Impr use less space for indexing
    Indexing, especially PIA, eats up an enormous amount of space both for RAM and the resulting files. Going to <Ref ...>line number list</Ref> and a corresponding data structure in the indexer fixes this almost completely.
End of list.