Text 8529, 84 rader
Skriven 2006-12-19 22:34:30 av Maurice Kinal (1:140/13.1)
   Kommentar till text 8525 av Ross Cassell (1:123/456)
Ärende: here again
==================
Hey Ross!

Dec 19 16:29 06, Ross Cassell wrote to Maurice Kinal:

 RC> Sheer number of files and file i/o.

I doubt that is an issue anymore but if it is then there isn't much in the
current pkt header that will assist speed and reliabity so it really isn't an
issue as to whether text or binary.  Once those are out of the way then the
only thing left is to scan for MSGID's if that is what your tosser uses for
dupechecking and that will usually fail if backwards compatibilty with editors
that don't even bother creating one of those in the first place or do it wrong,
for example the way Sychronet used to create them.

 RC> I presume you advocate a single text file for each message?

No but it would be handy if there was an unique delimiter to indicate the end
of each individual message within a mail archive.  At present it is a null
which reveals nothing since each message as well as both the pkt and msg
headers contain a whole mess of, few of which are really needed since much of
the info is 16-bit ints or a set number of bytes.

 RC> This is 
 RC> sort of like the original fido/opus message base, but even those 
 RC> files are binary in nature.

Fine but the bottomline is that in order to communicate universally on a
network then it only makes sense to have an accepted standard for networked
transfers.  At present that is a real can of worms especially considering the
differing transport agents being employed or planned to be employed coupled
with the method of displaying or transferring (offline messaging) to users on
any particular node or so-called BBS which varies greatly these days as to what
defines a true BBS.  Your newsgroup<->fido gateway is a good example.

 RC> The more messages there are in even that setup slows down scanning 
 RC> and tossing, like I said too much file i/o.

That is debatable.  Right now far too much of the input part is highly
questionable as to how much usefulness it is to any scanning and tossing
scheme.  For instance I wrote a Perl script that could noticably outperform hpt
wrt speed at packing messages even with all the 'proper' kludges to call it a
raw FTN pkt in play.  Unpacking is a different can of worms which it shouldn't
be but is.

 RC> I have been in here for quite some time, longer than Wayne as a 
 RC> matter of fact, just because I didnt post 12 messages a day to meet 
 RC> your criterium is meaningless.

As well as any criticism you have of Wayne.  For the mostpart he is far more
ontopic then you have been no matter what objections you personally have with
some of what he says within a message.  As for my opinion being meaningless I
could care less about that simply because I am right and you are wrong.  No big
deal as I have no issue with you or your sheeple and don't lose any sleep over
your current strategy.  Fido has been slipping all on it's own into obscurity
so I doubt this all matters in the grand scheme of things.

 RC> If you look back, it was Wayne who brought his baggage in here, 
 RC> although I suspect he fat fingered his keyboard.

Whatever.  Doesn't matter methinks.

 RC> Do you ever get to use your computer or is it still using you?

For the mostpart I am in control.  Some things, like the internet for instance,
I completely lack control over other then picking and choosing what I care to
care about.  I am happy that ftp still works and if it didn't I'd probably lose
the connection to the internet since most of it, especially the www, is of
little consequence to me personally and cost far too much to concern myself
over.

 RC>>> Except the version I wanted and looked for was 1.1, the version I
 RC>>> got via deb is 1.1.

 MK>> Okay.  You should be happy then.

 RC> So using the deb was ok?

Sure.  Is that what your Windows users use?  ;-)

Life is good,
Maurice

--- Msged/LNX 6.2.0
 * Origin: The Pointy Stick Society XXV - Almost there (1:140/13.1)