New classification and question on networking (was: Re: [Demexp-dev] Issues on classification)

[David MENTRE]

David MENTRE <address@hidden> writes:

> The remaning modules to write are thus:
>
>  - classification (only simple keywords and path matching)

Ok, I have updated the CVS with a very very short module (16 lines of
Caml, on this one I have more comments than code :) that does the
classification.

[ ... ]
>  - network messaging (the minimal required set)

I have a question on network messaging: what kind of encoding should we
used for messages on the network? I see two main choices:

 1. use an HTTP like encoding, i.e. basically ascii strings

 2. use an XDR like encoding (following corresponding RFC), i.e. fields
    of determined size with a typed content (unsigned 32 bits integer,
    string, ...)

I think option (2) is better because:

 - no need to parse strings of variable size, no need to take into
   account variable spaces or OS dependent end-of-line

 - XDR fields should be shorter, so better network efficiency

 - would be possible to have efficient encoding and decoding

Advantage of option (1) would allow to follow the habits of the Internet
(like HTTP, NTP, FTP, ...).

Asking a colleague at work, he would also advise for option (2).

What do you think of it? Any advice from previous experience?

Otherwise, I think I'll use the usual suspects: TCP socket, several
messages between the client and the server, TPKT messaging.


Yours,
d.
-- 
 David Mentré <address@hidden>
   http://www.linux-france.org/~dmentre/david-mentre-public-key.asc
 GnuPG key fingerprint: A7CD 7357 3EC4 1163 745B  7FD3 FB3E AD7C 2A18 BE9E