Text 753, 86 rader
Skriven 2007-03-24 08:30:23 av Mithgol the Webmaster (2:5063/88)
    Kommentar till text 746 av Scott Little (3:712/848)
Ärende: Fidonet URL syntax
==========================
And about 22:43 23 Mar 07 it was written from Scott Little to Mithgol the
Webmaster:

 SL>>> At least, for an echomail area.  It's fine for a file echo.

 MtW>> Define 'hierarchical resource' then. Because I just can't get it.

 SL> hierarchy (n): any system of persons or things ranked one above another.

 SL> 1) A file attachment is content not structure
 SL> 2) A file is not a logical subordinate of a /message/ echo.

I know that echomail area is intended to be a message echo. However, I can't
ignore the fact that UUE-encoded content is common nowadays. That's why files
are delivered not only in file echoes (fecho://) or by file requests (freq://),
but also in message echoes (area://) and by FAQ servers (faqserv://). I had to
accept this fact as an established, and to impletement an uniform hierarchical
structure of object-path though my proposal of a standard.

 SL> If there is to be a synthetic structure, a date based filter such as
 SL> /yyyy[/mm[/dd[/hh[/mm[/ss]]]]] would make more sense and be far more
 SL> useful. Regardless, once you go there, it's etched in stone.  In HTTP
 SL> land, the object field is a static handler, while variable content is
 SL> passed in the optional field, thus allowing the server to support any
 SL> number of schemes.

Great Scott!

We don't need any synthetic structure for the sake of structuralism of anything
ideal; the need of object-path is dictated by the reality. Once or twice a week
I see binary content sent through message echoes, UUE-encoded. That's why my
proposal of URLs has to accomodate this ablilities. I am not going to use the
standard of FGHi URLs to put a certain pressure upon those who send files over
message echoes, I am not going to invent a tool to bundle off the UUE encoders.
Who needs an URL standard that is not even able to address the files inside
echoes like Fido7.Sound.UUE, Star.Wars.UUE, inside some gamers' echoes with
additional levels, et cetera?

But if your feeling that a file is not a logical subordinate of a message echo
(and probably should not be sent over echomail) is strong, then you'd probably
be amused by the thought that the mere transition to hypertext is likely to
cause some cease in UUE traffic. Some people use UUE because they want their
binary files to be seen by the readers of the letter they embed UUE codes in.
(Posting the file somewhere else would cause some additional manual operations
necessary for the readers.) However, if it were a hypertext, than a hyperlink
would be a good replacement for these codes; a posted URL is instantly fetched
by the browser and the illustration is displayed along the message text. If it
works across the different sites, it would work in Fidonet as well.

That's why I am not going to place an additional restrictions and to introduce
an URL system that is able to designate files in file echoes by not in message
echoes. Let URLs to document an existing habits of Fidonet users, and let the
hypertext change them.

 MtW>> Do you mean that different echomail/fileechomail areas in different
 MtW>> FTN domains may have same names, so it should be possible to append
 MtW>> domain names to area names to prevent ambiguity?

 SL> Exactly.  And the same for zone numbers.  I don't know of any
 SL> implementations that do this though, but since you're writing a next-gen
 SL> specification I can only hope that one day, 20 years after they were
 SL> introduced, someone might finally implement domains properly ;)

 SL> Again, I'd expect domains only be applied locally, not included by
 SL> default. ie. a Fidonet reader would know what the domain is from the
 SL> current context, but if a link were to be exported to a
 SL> non-context-sensitive environment (eg. clipboard, or dragged to the
 SL> desktop as a shortcut) the domain would have to be embedded on the fly.

Well, now I am sure that I know what you mean. Though, however, I am still in
doubt about whether something like that is really needed. What is more common:
different echoes with the same name, or echoes that are gated across domains so
that the echo name persists? I probably should just introduce the same 'uplink'
optional parameter in 'area://' scheme that was already introduced in another
scheme ('areafix:') ÄÄ and with an additional note that the conforming Fidonet
browsers should prefer uplinks from the same domain, even if they can't access
the exact node designated in the optional parameter. Would you be satisfied
then?


With best Fidonet 2.0 regards,
Mithgol the Webmaster.                 [Real nodelisted name: Sergey Sokoloff]

... BETA: Bugs Exist, Try Again.                        (Bugzilla Quip System)
--- Just as the clouds have gone to sleep, Angels can be seen in heaven's keep
 * Origin: And the Soviets ÄÄ waist-deep in the snow ÄÄ marched in (2:5063/88)