OSDev.org

BI lazy wrote: Oh, sometimes, ranting helps for it loosens the knots sitting at crucial places in /dev/brain (thx pype).

Thx

By the way, what exactly do you mean with "statistics?" in terms of metadata? I 'm a bit confused about this part. It's ok, the os doesn't need to know about file types and sort of ... this is for special modules as far as I understand it. and these modules then do the actual searching/combining/page or line counting (something like a wc?)?

Kind of. The modules are on the system, and some of them put themselves on default-on display and some on default-off, and in the dirlist for any number of files, you can select a statistic (something like word-count or artist or byte-size or fragmentation or anything some program can calculate) and it'll be auto-calculated (or gotten from the cache) and displayed for all files supporting that statistic. And, if you sort on that statistic (which is a GUI-design thing) it just asks the OS to make an index based on that statistic and then displays the B-tree (yep, I'm going to add B-trees) on screen.

@pype: Joe Average is of course happy with stuff hidden away from him, but will it make sarah programatzki happy, if she can't delve into the systems depths? May be you are right, may be there are more of Joe Average than of Sarah Programatzki outta there.

Sarah doesn't have to worry, as Pype says. I'm also an XML fan, so the metadata probably will end up in xml, so Sarah can just hack away at the cache (if she feels like it), she can configure which handler should handle which files, configure default handlers, write her own (with a demo module being freely delivered, of course) so she can have her own statistics displayed, instead of the ones I feel necessary (gbye windows, hello Intelligent Display). Of course, these settings (& the modules) are put on your WW-space so you can log on to a different computer and instantly use your own statistics .

In some cases the statistic-name is unfortunate, for instance, I'd like to make a module that extracts all words that are not in the dictionary and makes a dir-wide index of that, so you can search for search terms in your files. Didn't work that out too well particularly, but at least I'd like to be able to search for "Intel" and not expect every document to store that in ASCII or unicode or whatever my system happens to like. Helps searching compressed or encrypted files...