Tuesday, July 03, 2007

which is written as 'rm' and indicates that a reduction with grammar rule m should be performed

Overhauling my search string parser at work right now. Currently, it only supports quoted exact phrase searching and a few magic prefixes. I'd like to add support for "OR", "NOT"/"-"/"!", and parentheses for Boolean trees of arbitrary depth.* Yesterday's Wikipedia crash course in parsing has given me some second thoughts about going for a Master's--wow, this stuff is complicated. Perhaps I'm just that guy who knows the standard libraries really well. Or the plucky comic relief.

Makes me wish I hadn't been joking when I suggested to my boss that we require searches to be in reverse Polish notation.


* It's somewhat more convoluted than this because of real-world constraints: single-character searches are prohibitively expensive and are tacked on to nearby words for exact phrase searches, and negated searches only work at the end of AND sets. (Yes, both are probably because I'm going about the problem the wrong way in general.)

4 comments:

Thirdmango said...

Happy Birthday!

azurerocket said...

So, I was poking around on your links and saw that your company is likely the one that was right across from my bedroom window. I hated when those guys stayed late, and they did it a lot.

azurerocket said...
This comment has been removed by the author.
azurerocket said...

Ah, yes, I was talking about Proton. I used to live on the third floor of an apartment building at 1801 N 950 W, so I guess it was some other company in the same complex or a little further down the street. I know the nursing home you're talking about, Cove Point or something. My bedroom window was exactly opposite the back windows of one of the buildings in the complex just south of it.