create new tag
, view all tags
Hmm, the diff showed the change from a bulleted item to a table nicely (but perhaps verbosely)--with no changes within the text of the item, it showed first the item as a bulleted item being deleted, then the item as a table being added. Just for kicks, I'm going to try changing it back to a bulleted item expecting that the change will be shown in a very similar manner. I may then also try a change from a table to a plain paragraph and other similar changes. I'll probably only report back if I see something surprising. (It may be a few days till I make all the changes while I wait for the one hour timeout (and I'm not sitting here with baited breath).) (Or is that bated breath?)

Darn, forgot about the 1 hour wait to force a revision, thus my notes about rev 1.2 below are misleading, as the original r1.2 I was talking about was conflated with the further revisions (the next two paragraphs) in the subsequent within one hour revision. (Aside: I hate the word conflate, but I think (without checking a dictionary) I might have used it correctly/appropriately in this case (but which is not always the case for the use of this word).)

Wow, the diff from 1.2 to 1.1 looks rather ugly from a technical point of view. Shows the entire original first paragraph being deleted, then added back after the (3?) new paragraphs were added. I wonder how this will look? The other change I'll make next time is to start making some of this a table, but this time around I'll make a bulleted list:

Diff Comment R1.2 R1.1 Ugly--entire 1st paragraph deleted and then readded--was it because I deleted the initial date?

I'm thinking about writing my own wiki-like thing, and thinking about doing it in Ruby (partly because the last time I tried Perl I had a great deal of difficulty and eventually gave up). I should admit that a lot of the problem was REs, and I'll probably have to deal with REs in Ruby as well. Other parts of the problem were the cryptic nature of some things (all the $something variables) and then the complication (to me) of the TIMTOWTDI philosophy. I looked at Python which didn't look too bad based on a good book I found (recorded somewhere), then heard Ruby was better and decided to look at it. Then digressed to considering Lisp (quite a bit) and Smalltalk. Now I'm back to thinking about Ruby again, even though it can be quite complicated and has some of the TIMTOWTDI flavor of Perl (with some of the tricky features (continuations) of Lisp.

Anyway, I'm starting to write down my requirements, and began to realize I needed to solidify my requirement for diffs, so I want to do some experimentation on this page.

The following is quoted from my gjots file containing the notes of my wiki-like thing requirements--it is included primarily to have some text to start looking at diffs "against":

Diffs. I was thinking that diffs wouldn't be too hard (other than finding a program that does "word diffs", but then I got to thinking (worrying) about something else--I'll need to do some testing/investigation to see if my worry is justified, but recent confusion over the diffs I see on TWiki hint that I may be justified in worrying. Initially I think it's simple, I want to see differences as deleted (struckthrough) or inserted (bold or underlined) words or phrases, and in the same "style" as the more recent version being compared (I think, or maybe an option to do it in the style of the older version being compared). But then I began to wonder, how will this work if, for example, (trying to pick one of the worst cases, without knowing for sure what that is) if one version has the same text in a table and the other has it in plain text (paragraphs)? I'll try to look at some TWiki diffs where that occurs and see how they handle it (and maybe start a page on Codev to discuss this and some related issues).

On the same subject (i.e., diffs), it seems clear that the "canonical" diff should be a diff on the "canonical" raw wiki text, i.e., based on (differences) in wiki marked up text rather than differences in HTML marked up text. This if for no other reason than to support things like reversion--we (I) want to revert to the original wiki markup version of a page, not something that is HTMLified.

Even so, that leaves two options--display the diff as raw (wiki marked up) text or attempt to render it based on the wiki markup converted to HTML. In fact there are a number of possibilities:

  • totally plain diff: strip the wiki markup off either before or after running the diff
  • totally raw diff: wiki marked up text displayed as wiki marked up text
  • HTML rendered diff, type 1: do the diff based on twiki marked up text, then HTMLify the diff and render
  • HTML rendered diff, type 2: HTMLify the individual versions before doing the diff, then do the diff and render (not sure whether there will be any (practical) difference between the two)

Finally (but maybe it should have been firstly), I should think about the (possible) purposes of a diff--I think there are more than one:

  • For a reader, to easily recognize the difference between a version of the page he might have read and the current version without rereading the entire page
  • For purposes of (things like) reversion (after spamming)
  • ?For purposes of the writer--I had a vague notion that there might be some difference for the writer (as compared to a reader) (and I'm fighting a headache), and I can't at the moment recall (or see) what that difference might be.

I might take a little time to look at some diffs on TWiki.

Edit | Attach | Watch | Print version | History: r5 < r4 < r3 < r2 < r1 | Backlinks | Raw View | Raw edit | More topic actions
Topic revision: r5 - 2005-02-15 - 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