[Info-vax] OpenVMS Modernization Development Tools on YouTube (eCube)

[johnwallace4 at yahoo.co.uk]

On Thursday, 18 September 2014 22:25:59 UTC+1, Johnny Billquist  wrote:
> On 2014-09-18 22:46, johnwallace4 at yahoo.co.uk wrote:
> 
> >
> 
> > ..
> 
> >
> 
> > "More seriously: the world isn't going back to the land of 7-bit ASCII text."
> 
> >
> 
> > All these band-aids apply largely BECAUSE the RFC world **hasn't moved
> 
> > on from** the land of 7-bit ASCII text, and in particular hasn't moved
> 
> > on from the teletype era when applications and protocols had to be
> 
> > designed to cope with network links that weren't 8bit clean. The RFC
> 
> > world hasn't moved on from leaving dealing with those little details to
> 
> > the upper layer applications, because it's "an application layer
> 
> > problem". Yeah right.
> 
> 
> 
> What kind of nonsense is this? 8 bit clean have nothing to do with this, 
> 
> and TCP/IP is always 8 bit clean. There are some protocols on top of 
> 
> this which talks 7-bit ascii, and of course those protocols will 
> 
> continue to talk 7-bit ascii. They would do so no matter what the 
> 
> underlying protocol looks like.
> 
> 
> 
> > Oh hang on, we had this discussion not far from here a few days ago.
> 
> > And the OSI world had it thirty years ago. And still the RFC world
> 
> > doesn't "get it". Ho hum.
> 
> 
> 
> Yeah, right. So tell me, how does OSI deal with Unicode? And how does 
> 
> your newsreader that talks OSI deal with transferring both text and a 
> 
> binary file in the same message? Or actually, how does news work over 
> 
> OSI in the first place?
> 
> 
> 
> Oh? You mean the OSI model have not even thought about the problem yet.
> 
> 
> 
> Come back in another 30 years, when I believe the OSI model will be even 
> 
> more relevant than it is today.
> 
> 
> 
> 	Johnny

"There are some protocols on top of this which talks 7-bit ascii"

An application based on a protocol that cares whether it has an
8bit clean network link underneath is not a sensible modern
application, OSI or otherwise. Such an application is almost
certainly a relic from the teletype era.

Time to move on, whether the underlying layers are RFC-centric or
OSI-centric.

Yes of course IP is 8bit clean. Regardless, everything above IP is
*still* "an application problem".


"how does OSI deal with Unicode?"

Unicode *IS* inherently and unavoidably an upper layer thing
(application and presentation layers, 7 and 6). The layers below
neither know nor care about Unicode.

Various aspects of the OSI model do actually provide architected
ways for two or more co-operating applications to negotiate
various layers' parameters. In the case in question, they' have
an architected discussion about what character representation to
use.

It's not totally trivial, but from memory if both ends do Unicode,
you get a Unicode exchange if Unicode is what they want. If they
don't both do Unicode but can both do (say) DECMCS they may fall
back to DECMCS. The interconversion is then handled not by the
application itself, but in a lower layer.

A more obvious example would be an EBCDIC-based application talking
to an ASCII-based application. The OSI presentation layer handles
that magically in an architected way.

How does the RFC world handle that kind of presentation layer thing?
It doesn't, as usual "it's an application issue".


"how does your newsreader that talks OSI deal with transferring both
text and a binary file in the same message?"

In a sensible world, it would likely do so in a remarkably similar way
to the way DEC's own Compound Document Architecture applications worked.
Did you ever get a chance to play with CDA? Not many people did, which
was a shame. And then MS turned it into OLE and duly wrecked it.

Or, independent of CDA itself, although conceptually not a million
miles away, there were the designed-in (not bandaided-on) vendor
independent compound document features inherently part of X.400.

But a technically superior architecture does not always win in the
market. Hence here we are thirty years later still struggling with
relics from the teletype era.