home *** CD-ROM | disk | FTP | other *** search
Text File | 1998-08-01 | 6.2 KB | 176 lines | [TEXT/ttxt] |
- RELEASE NOTES FOR A SORT OF A KIND...
-
-
-
- Introductory chatter...
-
- Here's one you'll use every day (grin).
-
- This is a sort, a barely-adequate nevertheless robust kind of a
- sort. It's A Sort of a Kind (hereafter "aSoaK" for brevity's sake).
-
- It's robust in that it will sort arbitrarily large files. It's
- barely-adequate because it implements only the bare minimum of
- features I envision for a useful sorting utility. I produced it
- this way, in this form, because I, personally, need something now,
- and because I don't have time to do more than this right now.
-
- In fact, during testing, I hit upon a much faster scheme for doing
- the same job, but I don't have time to go back and rewrite, so this
- is shipping the way it is. If you're feeling like a poor relation,
- don't. This is "barely-adequate" only in contrast to what I _wish_
- I had time to do. It's better than anything else I've seen -- and
- my natural preference is to make do with someone else's tools
- rather than write my own. In particular, aSoaK succeeds where (for
- example) the sort in Word goes south, and it sorts to a more
- reasonable order. Moreover, my own results (one tester reports
- contrary evidence) indicate that it is about 3 times faster than
- Word on larger files (it's amazingly faster on small files). So:
- while I can imagine a much better sort in terms of both features
- and performance, this will (and will have to) do for now...
-
-
-
- System stuff...
-
- IMPORTANT: aSoaK is System 7 or above ONLY. Its only interface is
- Drag and Drop, so you cannot use it with earlier Systems. If you
- double-click on it from a System 6 machine, you will get an error
- message and the software will quit gracefully.
-
- To use aSoaK _with_ System 7 or above, simply select the files you
- want to process and drag them on the program's icon or an alias of
- it. New files with the extension ".SRT" will be created, and your
- original source files will remain unaltered.
-
-
-
- Usage notes...
-
- A Sort of a Kind will sort arbitrarily large text files. File
- size is limited only to allocated RAM, with the reliable factor
- being file-size + 128K. If you get a message that your file is too
- large, increase aSoaK's memory allocation like this:
-
- 1. Get Info on the file. Note its size in kilobytes.
-
- 2. Add 128K to the size of the file.
-
- 3. Get Info on aSoaK and change the figure in the Preferred Size box
- in the lower right-hand corner to the value you came up with in step
- 2.
-
- If you ever actually do get a file that's larger than all the
- memory you can give aSoaK, drop me a note and I'll see if I can
- come up with something.
-
- Practically speaking, the length of any one paragraph must be less
- than 16,000 bytes, and warning for exceeding this limit is built
- in. The chances of hitting this wall are even slimmer than having a
- file too large to fit in memory, I expect.
-
- The default sort order is strict ASCII (case is sensitive, and
- accented characters are not equated to their unaccented forms). In
- addition, there are three modifier keys that you can use at Drag &
- Drop time to influence the sort:
-
- 1. Shift-D&D induces a (quasi-)lexical sort. Accented characters
- are equated to their unaccented equivalents, and case is
- (semi-)ignored by a double-weighting scheme. In other words, case
- (ignoring accents) is taken into account _only_ where the
- characters are exactly equal (ignoring accents) when case is
- ignored. Like this:
-
- Case-sensitive: M < N < m
-
- Case-insensitive: m = M < N
-
- Double-weighting: M < m < N
-
- With a case-insensitive sort, there is no prohibition against "m"
- coming before "M". Double-weighting cures that ill. Take note,
- however, that accented characters are double-weighted _only_ as to
- capitalization. Accented words are not ghettoized nor even
- (semi-)ghettoized for being accented. If you Shift-D&D the enclosed
- file called "Alice & Bill Toy With Names", you'll see how
- (non-ghettozing) double-weighting works.
-
- 2. Command-D&D simply yields the sort in reverse order.
-
- 3. Option-D&D causes _exact_ duplicate lines to be shown only once.
- Case and accents are _honored_ when deciding if a line is a
- duplicate.
-
- These three modifier keys can be "stacked". That is, you can hold
- down more than one, and each will be honored. So, for example,
- Shift-Option-D&D will produce a lexical sort with exact duplicates
- shown only once.
-
- If you have used any of the modifier keys, the progress window will
- tell you which ones.
-
- If you make a mistake or if you're tired of waiting for aSoaK to
- finish, holding down Command-Period will abort the process.
-
-
-
- A Sort of a Kind in Real Life...
-
- aSoaK and Word: Word sorts by means of a (to my mind) excessively
- complicated mathematical weighting scheme. This is why it produces
- such odd results so slowly and why it breaks down with large or
- hirsute files. aSoaK's lexical sort (Shift-D&D) produces what Word
- is striving for, faster and cleaner, without the size and
- complexity limitations.
-
- aSoaK and Idiot Randomizer: Idiot Randomizer was built for Word's
- (goofy) way of sorting, so aSoaK's results will not be identical.
- They will, however, achieve the same end, which is why I didn't
- build randomizing into aSoaK.
-
- aSoaK and FontFischer: It was FontFischer that made me realize that
- I couldn't make do with what I had. Word chokes on large
- FontFischer files, and I have need to make a _big_ font book.
-
- aSoaK and Torquemada: Torquemada the Inquisitor saves me a lot of
- time, more than you might ever guess. My idealized sort would
- permit you to, for example, weight the fields of database report
- records, so that a report like this:
-
- "Greg","Swann","70640,1574"
- "Shane","Stanley","100033,317"
- "Kip","Shaw","72320,1301"
-
- could be sorted on the _last_ name, rather than the first. Users of
- Xdata will see the utility of this kind of power at once. But: we
- ain't there yet, and I won't have time for IdealSort for quite some
- while. However: here's where Torque saves me (and you) a lot of
- time. This set:
-
- ^?","^*","^~^p
- ^*|^?","^*","^~^p
-
- will prepend a sort key consisting of the last name field to each
- record. You would run the resulting TQM file on aSoaK, then use
- this set:
-
- ^?|^*^p
- ^*^p
-
- to strip out the sort key. All sorts of analogous stunts can be
- pulled, using Torquemada or other text-processing utilities.
-
-
-
- Very Best,
-
- Greg Swann
-
- gswann@kagi.com
- gswann@primenet.com
-
- USPS: 3608 West Cochise Drive
- Phoenix, AZ 85051
-
- 8/1/98
-