On 11/14/2000 9:12 AM, "Remo Del Bello" <[EMAIL PROTECTED]> wrote:
> on 11/14/2000 12:29 AM, Barry Wainwright deftly typed out:
>
>> At 5:13 pm -0500 13/11/00, j.d. o'grady wrote:
>>
>>> as a 5+ year eudora user I would humbly suggest that you index the message
>>> body field. in eudora you can do an advanced find (cmd-opt-F) and find
>>> anything - FAST! as a high-volume emailer (300+ inbound/day) I can tell you
>>> that this is the most disappointing aspect of entourage.
>>
>> But, indexing is not necessary. Eudora doesn't index the message body
>> and it's searches are much, much faster than OE's or E's as Jason
>> pointed out.
>>
>> It should be possible to improve the searching without having to
>> resort to indexing every word which would, I suspect, have a
>> tremendous hit on receiving/filing mail (as it does in Filemaker -
>> try comparing an import into a text field when it is/isn't indexed!).
>
> I agree with Barry. My quick, unofficial test indicates that both BBEdit and
> (according to Barry and Jason) Eudora's text searching is infinitely faster
> than OE/Entourage's. And that is without indexing in either. Perhaps the
> search routines need to be re-written if they can't be optimized to the
> point of being useful without implementing indexing. There are other
> examples of software that is VERY fast at searching text (Grep, Perl) and I
> would hope that with all of the resources available at Microsoft that there
> isn't a really efficient text search routine around somewhere that could be
> borrowed for Entourage.
>
> I, for one, would love the improvement in search speed that indexing the
> message body would provide, but I don't look forward to the performance hit
> the application would take at either the receiving/filing stage, or, as Dan
> suggested, in the background when idle. And I don't believe that, with a
> good search routine, that it would be necessary. It would be something that
> could be implemented in time for the next update, however, whereas a
> complete rewrite of the search routine is not.
There are a couple of issues that slow us down. One is that we store the
original source of the message only. I believe that Eudora stores a decoded
version of the message. The second is that we support Unicode and multiple
languages. So, for each message body we search, we have to decode the
message as Unicode and search the Unicode.
There are certainly optimizations possible in this process, especially when
searching for ascii text.
Dan
--
To unsubscribe: <mailto:[EMAIL PROTECTED]>
To search the archives:
<http://www.mail-archive.com/entourage-talk%40lists.boingo.com/>
