Tags:
create new tag
, view all tags
May start some diary like pages for a few reasons, one being that much as I like wikis, I sometimes think that it would be nice to have some more blog like features (as well as more email features as I may have mentioned at other times).

Maybe I should call this page Rhk2003November and (potentially) have headings for any day that I make an entry. We'll try that.

See:

Contents

10th, Monday

Wikis, Offline editing with Nedit, Nedit folding macros, askSam for Linux

Something about the current status (of whatever):

  • I've been using my poorman's askSam for Linux (the first versions being TWikiText files) for probably 6 to 8 weeks
  • Probably a month ago my desire to have "per heading" folding macros for offline editing was rekindled
  • After a bunch of false starts, I finally believe I have the logic worked out, for a number of alternatives, including well-formed and ill-formed outlines
  • Because of the variable length of TWiki headings ("---+" for Level 1, "---++++++" for Level 6), and the limitations of the Nedit macro language and REs (just theirs or all REs?) the coding is not as straightforward as I'd like. (I put an RFC (should have perhaps been an RFEC??) on Codev outlining the problem, with some notes on the Meatball Wiki as well.)
  • I got sidetracked into looking at CVS and RCS again, with the objective of simply learning how to put one file under revision control — I was pleased to find that the man page (for both) now make sense to me, and that the info page for CVS is actually different than the man page, and seems quite useful as a learning tool.
  • And, then I started on a quest to get plain text copies of the info files so that I could do my thing with respect to copying, condensing, and rewording such documentation into my own words.
  • I found that:
    • info files can be viewed in a plain text editor (but do have the typical man page problems that make it difficult to edit (indented paragraphs, etc.)
    • info files can be developed and derived from the "typical" Linux docbook document development tools
    • I was reminded of the LDP's wiki like thingie to edit LDP documents, and read the (short) HOWTO
    • that left me with some questions that I started doing some googling to answer (will list some below or on another page)
    • looking at some of the syntax for the wiki editing tool (one page by somebody describes Meatball Wiki as the canonical source of wiki syntax — I want to find those pages again for whatever good they can do me, as well as suggesting that if he gives credit to David Merrill for "wikitext", Ward Cunningham (and perhaps others) should get some credit for developing wiki.
  • I'm recognizing more of a problem because now I have many of my recent (Linux era) notes split between WikiLearn and the offline askSam thingie (which I do hope to find a semi-automatic way of exporting to TWiki).
  • Using info reminded me that it could possibly be a reasonable candidate for an all in one askSam for Linux thingie (whatever it turns out to be)
  • Anyway, considering all of the above (and possibly some other things not cited above), I think it may be an appropriate time to relook at wikis and alternatives, which, if you haven't already guessed, I've started and will spend some more time on

Some questions about the LDP's wiki like editing tool

It is clear (from what I've read) that you can go "wikitext" --> docbook, can you also go docbook --> "wikitext" so as to allow you to edit existing LDP documents that were not originally created in the wiki tool. (Some other keywords for my search: David Merrill,

Notes during my research today

Oops, I think I'll put them on the askSam for Linux tool, it's easier to reclassify the notes after after creating them (by adding or changing the title and resorting), plus I don't have to go through the frustrating edit / preview / save / re-edit cycle (I just press <ctrl>s in Nedit (or Notepad, or Wordperfect) and have instantly saved my work and ready for additional editing. And, even if I'd want to search for a particular page to put a particular note on, it's just a matter of collapsing the "outline" and scrolling. Hmm, maybe I'm finally articulating some of the other frustrations of TWiki that have annoyed me.

15th, Saturday

Ahh, good, I notice that I did mention the idea that "it's easier to reclassify the notes after after creating them" (than searching for (or creating) the right record to add them to in the first place). Reason I'm glad (not that it's a big deal, but I thought that was an independent realization by myself, then, sometime (late) this week I found an explicit mention of that on the Chandler wiki (Big Ideas).

More on Chandler: Sometime this week, I got, AFAICR, my first WebNotify email from the Chandler wiki. So I went to explore. It seems Chandler is still chugging along (although in places it seems like they are so much into research that they may never get a product out the door (and then I read some (all?) of their biweekly "all hands" meeting summaries and maybe some other stuff that lead me to believe they have a happy balance between research and actually building a project, and, when they get something out the door it could be very interesting. I need to reread the BigIdeas page (and some others, IIRC, something like the Grand Reunification Theory of Virtual Files) — some interesting ideas there about classifying items in multiple (orthogonal?) categories, the idea of leveraging an individual's memory by (IIR/UC) recording the stuff he has trouble remembering but taking advantage of the stuff he can remember. (Interesting little brainstorming scenario about instructions for the babysitter.)

Anyway, one of my points here (just for my own personal satisfaction) is the belief that I came up with the "it's easier to reclassify the notes after after creating them" independently, and before I read it there. Amazing coincidence though, that you did it the same week? Yes, but I still think I did it independently. I think some (inconclusive) supporting evidence would be finding out when Chandler got their WebNotify email working. (Because that's my belief, that it wasn't working, and then Mitch Kapor asked about it (because, IIRC, from the date on his "Main." page, he may have just registered on it around November 6 (, 2003), then signed up for WebNotify, then asked why he wasn't getting email notification, so then somebody dug into things and fixed it. (An alternate explanation, not likely, but not altogether unbelievable is that my local Procmail setup started doing something different with their emails, so they ended up in my "main" inbox instead of either getting discarded (accidentally) or ending up in some obscure inbox that I haven't read. Anyway, I guess if they did have a problem with email WebNotify, and fixed it sometime after 10 Nov, I'd have some (circumstantial?) support for my "claim" to independent development.

(_So what if you do? Well, maybe I have a slight claim to a great mind as in great mind's think alike? wink And/or more support for my theories that necessity is the mother of invention (my difficulties with entering data into TWiki triggering first the development of the (offline) askSam for Linux thingie, and then my current habit of entering items in blank records at the top of the file, then adding a meaningful (I hope) title, and finally sorting the records later (collapse to Level 2, Shell -> Sort).

Ok, enough of that! I originally planned to come here today to update status on some of the things I mentioned on the 10th:

  • WikiText may still exist, but David Merrill hasn't responded to the email I cc'd to him with questions including how to try it out, nor has anyone else on the list. Martin Wheeler did respond saying he used it a few years ago, and then stopped when he (seemingly?) lost access to it. I also read somewhere some comment about David possibly being sick again. So, for now (and for more reasons that might be covered below (and because I don't particularly like the WikiText syntax nor do I want to learn any new syntax)), I'm not going to pursue (the use of) WikiText (some of what they may have done may be useful examples for some of the things I would like to pursue). PS: The tools that exist convert WikiText to DocBook, not vice versa. The WikiText "tool" (I assume there is such a thing) apparently let's you deal with existing DocBook documents by displaying the DocBook markup (and allowing you to edit it?). Anyway, the more important capability for me is DocBook to WikiText, and they don't seem to have that translation capability. _Ok, a few days ago (yesterday) someone on the list (David Lawyer) said they called David Merrill — Lampadas seems to be merging into plone, a beta might be available in about 6 months (i.e., June, 2004??).
  • I re-realized that AbiWord has both DocBook and Wikipedia import/export filters. That and other things make me think that AbiWord could (now or someday) be WYSIWYG editor for DocBook (if it's not done already, define appropriate styles (title, whatever) that include the DocBook markup). And, because of the way AbiWord works (IIUC), translating "foreign" documents directly to and from the piece table, ... (more than I want to explain right now, and maybe I've mentioned enough, if not here, elsewhere including some emails to the ldp-discuss list).

Bottom line: I think pursuing some means to convert DocBook to TWiki markup is something I want to put on my Wish/ToDo list, even recognizing that such translation is lossy (there is more information in DocBook than would normally be stored in a plain text file with TWiki markup (I can expound on that some other time, or find a page to refer someone to, but just as a quick example, in DocBook, I presume that portions of a document can be marked as code, programming, or whatever, while TWiki markup would simply show that as == (or <pre>, <verbatim> and not retain the information that the portion is code).

I can see three four paths to achieving that:

  • a Nedit macro
  • an awk, sed, or Perl script
  • an AbiWord export filter for TWiki (and perhaps refining or maintaining the DocBook import filter), and then, for completeness, dealing with the reverse path (TWiki import, DocBook export)
  • hmm, how about a DocBook stylesheet? sounds like it should work

All three of the above would be educational for me. The awk, sed, or Perl script might be useful in terms of (re-learning) how to create a script in one of those languages, but since at least an awk or sed script would primarily use REs (IIRC), it may be a good first step to develop those RE's in Nedit, then translate the script to awk, sed, or Perl. (Unless the RE's get cumbersome and I find some alternate approach that requires more "if then else" type flow control.)

Doing the AbiWord filter might get me "in" to the AbiWord codebase and vocabulary.

So, I won't make a decision right now, I might even just grab (copy and paste) a text copy of the CVS info pages to start my copy, condense, convert to my own words process for CVS (or, did I find and download the DocBook version — if so, I might fool around with converting it "semi-automatically" in Nedit and keeping a record of my steps (I could do a keystroke recording of the entire session (is there a limit on how long that can be (there is a limit on my patience in sorting through some long recorded macro wink ), or just keep a manual record of each find and replace I do.

Was there anything else I wanted to record here today? I mentioned finding Chandler again, I also found the CollabWiki: Pattern Repository and find it very interesting, but I'll record more about it in my offline notes. Oops, I should at least provide one link to each, so ... done.

I'm back (don't know if it's been over an hour to cause this to be recorded as a revision, I'll know soon, though).

So, what is my next priority? Possibilities:

Copied some of the following to ComputerThingsToAccomplish — update it there.

  • Develop mechanism to move askSam for Linux stuff onto WikiLearn?
  • Learn CVS / RCS (and record learning using "brute force" copying of CVS info)
  • Script for above (convert CVS info page to text or TWiki markup)
  • keyboard macros (digression: dig into xnee further, write to developers (IIRC, I joined the mail list))
  • AbiWord collapsible outlining
  • AbiWord export filter for TWiki markup
  • Finish NmTWikiFoldSingleHeading (for either one of the simple cases (constrain the outline), or for an arbitrary outline after translating TWiki heading markup to "---+2" style (and vice versa)

Oops, almost forgot:

  • Move WikiLearn to new site
    • Consider OWiki vs. TWiki (wait for next release?)
    • Redo templates (convert from old style to modular)
  • Learn Python (then C and C++?)
  • Learn programming infrastructure tools (CVS, make, GCC, GDB, Bonzai, ...)

Hmm:

  • Maybe some of the above are long term "stragegic" (?) goals (Learn programming infrastructure) while some on the list below (Learn CVS) are the subgoals (or "tactical" goals) — I might want to have separate lists for the strategic and tactical goals
  • and move the whole list of goals and priorities to another page (do I already have one? — maybe the name should be) Prioritized Goals?

Tentative prioritization:

  1. keyboard macros subset: dig into xnee further, write to developers (IIRC, I joined the mail list)) — write to the mail list (seems an easy enough starter task)
  2. Script for above (convert CVS info page to text or TWiki markup) — spend a little time investigating DocBook, find the DocBook file for CVS info, try a semiautomatic translation (find and replace in Nedit)
  3. Learn CVS / RCS (and record learning using "brute force" copying of CVS info)
  4. Finish NmTWikiFoldSingleHeading (for either one of the simple cases (constrain the outline), or for an arbitrary outline after translating TWiki heading markup to "---+2" style (and vice versa)
  5. Develop mechanism to move askSam for Linux stuff onto WikiLearn? wait awhile, get more experience to see what kind of content ends up on the offline file
  6. Move WikiLearn to new site
    • Consider OWiki vs. TWiki (wait for next release?)
    • Redo templates (convert from old style to modular)
  7. AbiWord export filter for TWiki markup
  8. AbiWord collapsible outlining
  9. keyboard macros (modify X server)

18th, Tuesday

Made another mistake — had an idea yesterday or the day before and failed to record it (and a headache which is no real excuse) — had to do with, oops, now it's coming back, sort of:

Adding more links to a TWiki page to improve navigation within the page (OK, that is the idea, but you're missing the reason — what were you planning to accomplish? Well, it sort of tied in with folding, or an alternative to folding ... darn!

Hmm, maybe it had to do with simply taking some of my offline askSam for Linux pages and simply copying them to a single TWiki page, but then needing a way to navigate between items? (Or the more recent recognition that there is no good inter-item navigation on the offline thingie because there are no (workable) links while in, say, Nedit.

(Which led me to think about searching, navigation gets worse as the offline file gets bigger (the elevator bar becomes tiny, for one thing, and searches have to be based on more than one term, which I might be able to do in Nedit as a regular expression if I can limit the search to "records" containing more than one term, which I might be able to do while the outline is collapsed, because then each record is one long line (in which other line breaks (\n) have been replaced by \f (form feeds).)

But, I'm still dancing around what my idea was the other day. Oh, well, I'll save this, I suppose it will come back to me and this time I'll record it (on paper if not here).

Going to test some searching for multiple terms.

I'm back, OK, that's fairly easy, but also a little painful.

Much of this should be moved to the askSam for Linux page.

In general, you can search for a regular expression like: <term1>.*<term2>, but actually, you have to search for the terms in either order (unless I'm missing something). To search for three terms (and'ed), you'd have to search all permutations of the search string (six). It might be worthwhile at some point to create a macro to do something like.

The search is limited to one "line" (but does "cross" \f's), and I was going to say that I don't see a good way to do proximity searching, but I do now: By setting the outline to different levels of collapse, I do have some control over how close together the terms must be, i.e., if the document is fully expanded, the search finds the terms all in one paragraph (line). If collapsed to say Level 2, it finds the terms all in one Level 2 section. (I should rephrase the previous to something like, "the search is successful if all the terms are found in" one paragraph, Level 2 section, or whatever.)

So, as often, it is not a good idea to assume that I knew everything there was to know about "my" poor man's askSam for Linux when I first thought of it — using it and experimenting with has led to the discovery of more capabilities than I originally hoped for. (Of course, it could have gone the other way — the real point I'm trying to make (not 100% sure it's valid) is that documenting it to the nth degree (and in "perfect" style) within a few days of thinking of it would have led to a need to revise the documentation fairly frequently, resulting in ugliness <==> frustration. On the other hand, some documentation is appropriate, but keep it bullet type items and what I'm afraid I'll forget._

Hmm, am I now (implicitly) touching on one of the things I was thinking about the other day — a need to include the same information in more than one askSam for Linux record? (Because I'm recording this here, but it really belongs on the askSam for Linux page, but I want some blog like recording of sequence of events?)

Other recent innovations in askSam for Linux (I need a better (shorter) name — kaskSam in the kde tradition?):

  • recording keywords on a blank line below the heading

Other problems recognized recently in askSam for Linux:

  • no linking while offline (in Nedit, for example)
  • the heading names I'm using (like ---++ File Transfer: SSH) are more like classifications than real page "titles", but maybe that's a passing problem, maybe I should concentrate on the heading becoming a more useful page title like thing (and possibly leaving the classification terms for another line??)

Contributors

  • () RandyKramer - 10 Nov 2003
  • If you edit this page: add your name here; move this to the next line; and if you've used a comment marker (your initials in parenthesis), include it before your WikiName.

Revision Comment

  • %DATE% —

Page Ratings

Edit | Attach | Watch | Print version | History: r5 < r4 < r3 < r2 < r1 | Backlinks | Raw View | Raw edit | More topic actions
Topic revision: r5 - 2003-12-21 - RandyKramer
 
  • Learn about TWiki  
  • Download TWiki
This site is powered by the TWiki collaboration platform Powered by PerlCopyright 1999-2017 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding WikiLearn? WebBottomBar">Send feedback
See TWiki's New Look