|
|
|---|---|
.JPG)
Interesting Words (Home); Proof-of-Concept ; The Interesting Words Engine ; Why is the Interesting Words Engine so Important?
Proof-of-Concept
The Proof of Concept was achieved by taking an existing long text in machine-readable form, a text which included a decent index, and essaying with types of rules, spitting out a list of interesting Words for comparison to the original index.
A set of developer-inspired rules and their user-defined values was soon achieved and gave a subjectively good facsimile of the supplied Index.
All that remained was to receive the rules and values from a simple table in a Word document, and to build a concordance tables from the result.
Feeding the concordance table to Word is a one-line statement in Word VBA!
From which, of course, we obtained our index.
In the space of less than a day we were equipped with a generic adaptable tool to create an Index at the foot of any document.
Should we wish, the index could be tailored to the TYPE of document (medical, pharmaceutical, aerospace etc) by tweaking the table of rules.
Now, for each one-day course, our workbook had an index, and my client looked much more professional to their clients!
Is it Useful?
Show the index to any technical writer and they will turn up their nose at it. (It's what technical writers do with any document they have not treated!). They are right. On close inspection an automated index contains several glaring errors, items that have slipped into the index that should not appear, and items conspicuous, as they say in the trade, by their absence.
None the less, when you are sitting at a classroom desk after lunch and can't remember where "Applications" was discussed, finding that phrase in the index, even when it is surrounded by useless words, is all that matters.
For the casual user, as long as 95% of the target words are present, that's all that matters. That 10% of the index is fluff matters not one whit. (Why did I use 10% and not 5%?)
For the serious user (Technical Writer), the automatically-generated index relieves the burden of 95% of the work, freeing the Technical Write up to get on with the job of the truly tricky words and phrases.
If the Technical Writer is processing documents for the host company, then time will be spent devising subtle versions of the rules tables for that particular industry. Indeed Indexr's adaptability means that one can drive the process with any one from a set of customized rules tables.
And Then One Day …
A prospect overseas asked for a sampler, and commented that the Indxr produced only a plain index. No chance of putting in multiple indexes, or "See also:" and "Refer to:" entries. In my usual style I snorted that THAT was beyond the power of Microsoft Word, but took a closer look just in case.
Sure enough, Word supports 27 indexes (one unnamed and 26 identified by a letter of the alphabet) and Word supports an unlimited number of types of reference phrases.
Back to the drawing board and out came Version 2.
I mentioned "phrases" earlier, and implemented two more data sources for the concordance table: KWIC and leader phrases.
KWIC phrases
Key Word In Context was a respectable indexing tool made popular by IBM back in the 1960s or thereabouts. Key words (in our case candidates for indexing) are presented within the context of their local text, and the text aligned so that the keyword appears centrally.
Above is represented a small example from a single issue of BYTE Magazine may 1984. The article from "Trump's card" has provided several interesting words, and for each word a copy of the line has been issued with that word rotated to column 40 (Yes, this was a primitive 80-column punched card version I wrote!)
For each Interesting Word we discover, we can lift and rotate the text and use it as a KWIC phrase.
Leader Phrases
A leader phrase is a phrase that is lead by an Interesting Word.
Above is represented a small example from an index run on this document while it was a work-in-progress. We have located several interesting words, and for each word a copy of the line has been issued with that word at the left-hand end of the phrase.
For each Interesting Word we discover, we can lift the trailing text and use it as a Leader phrase.
We now have three contributing lists – Words, KWIC phrases and Leader phrases – which we can mix-and-match in the Indexer.
It becomes a powerful tool very quickly.
The Indxr is acclaimed by its users. "Why didn't Microsoft think of that?".
Heh heh.
Loading
Toronto and Mississauga, Thursday, April 14, 2011 11:51 AM
Copyright © 1996-2011 Chris Greaves. All Rights Reserved.