Re: Newsreader on-off line (RE: Mix of Ideas ....)

[Gianluca Montecchi <gm518464 silab dsi unimi it>]

On Tue, 10 Feb 1998, Guillermo S. Romero / unnamed / Familia Romero wrote:

> >>         - the possibilty to search for something in the news server (don't
> >> know if this is possible) and download the articles that matter or even the
> >> threads that include those articles.
> >I am not sure that this can be a good choice forwho have a connection via
> >modem to internet (and pay bill phone based on the time)
> 
> Well, with the propossed way (copied from FreeAgent, and then improved) it
> can be done:
> - Connect and download the groups list (or refresh if you have an old one).
> Hang connection.
> - Search and mark interesting groups (by hand or by computer search function).
> - Connect again, dowload all headers of groups those which seem interesting.
> Hang again.
> - Search and mark interesting messages (by hand or by computer search function).
> - Connect again, download all bodies that have interesting headers. Hang again.
> - Read and reply when & what you want.
> - Connect (again?), send your replies and refresh you database (infinite
> loop or only keep reading some groups, called "suscribed to", refreshing
> only headers for that groups).

That is the better way for who pay a phone bill based on the time (like
here in Italy and in Spain as I understand).

> The search engine is better than FreeAgent, but the group->header->bodies
> way of work is a direct copy (copyright problems? are "adding 1+1" or
> "alphabet" copyrighted?). =;]

Skim use this method, and as I know have not copyright problems.

> All these ideas end in:
> Should a server and client app be coded separated? The server will manage
> queues and connections, and the client will show things to the user and make
> request to the server (the client must be capable of doing requests and
> retrievals).

Yes, it is possible to code separatly server and client. But I suspect
that the client must be like a system daemon or am I wrong ? 

> I say this cos I doubt if normal clients will be capable of realizing that
> unavaliable requests are queued and they have wait some time to get them
> downloaded, or they will directly say "not in server, forget about this
> article".

All the two things: if the server works with many clients (multiuser
situation) it put the request is a queque and every user must wait that
the request is processed; if the server works in monouser situation it can
collect a number of requests, than connect and do all the jobs.

Gianluca

    Member of the Pluto Linux Users Group 
  Avaible e-mail address:                                
  gm518464@silab.dsi.unimi.it (-) gianluca_montecchi@rcm.inet.it 
  Web page: http://www.silab.dsi.unimi.it/~gm518464
  For Linux questions please mail to linux@saturn.media.dsi.unimi.it