On Fri, Sep 15, 2006 at 09:11:04AM -0700, Jim Busser wrote:

> Is the issue that an effective search process must easily figure out  
> the method needed to look inside a particular object,
No, that's not the issue. It would be the issue, however,
were we considering such searching through blobs which we
are not.

> but plain text  
> file content does not itself say, for example, "I am an ASCII text  
> file"?
That, however, is correct. That's why I suggested converting
the "plain text" to unicode encoded text and store it in a
unicode encoded file with the proper markers saying "I am a
utf8 little endian text file". It is still text but it's
sufficiently self-descriptive.

> Does GNUmed have --- or is it in future intended to have --- a search  
> process that, based on start of file markers inside objects, can  
> easily determine a method to do e.g. boolean searches inside the  
> object(s), facilitated by some kind of re-usable open source plug-in?
No and no.

The docs viewer does work in a similar way, however. It
exports the blob into a file and uses some OS level
heuristics to determine the MIME type of the exported file
and runs the viewer associated with that mime type.

Karsten
-- 
GPG key ID E4071346 @ wwwkeys.pgp.net
E167 67FD A291 2BEA 73BD  4537 78B9 A9F9 E407 1346


_______________________________________________
Gnumed-devel mailing list
[email protected]
http://lists.gnu.org/mailman/listinfo/gnumed-devel

Reply via email to