Good basis. I'll insert <HL> comments in your text, below.
Harry Lewis
IBM Printing Systems
harryl@us.ibm.com
Michael Crawford <mcrawford@iready.com> on 03/25/99 03:32:49 PM
To: Harry Lewis/Boulder/IBM, Michael Crawford <mcrawford@iready.com>
cc: "'Richard Shockey'" <rshockey@ix.netcom.com>, ifx@pwg.org
Subject: RE: Word docs
--0__=oY3Zm2toYZe8fqdRVkuGGdAEz1QoQ2pC8hsoIPkkdnFl9fBBoaDpfjpI
Content-type: text/plain; charset=windows-1257
Content-Disposition: inline
Content-transfer-encoding: quoted-printable
I understand the confusion...I wish I wasn't confused myself, but then
again, wishing doesn't make it so (red face).
It might be worthwhile setting the stage so to speak.
Okay, this is a fax centric view. First a fax machine operation.
1. User takes out of box, plugs in power and hooks up phone cable to p=
hone
plug.
2. User configures name to appear on faxes it sends
3. User configures phone number to appear on faxes it sends
4. Users configures fax to auto answer when someone calls (probably a
default condition)
Ready to Go!
5. Someone elsewhere decides to send to fax to our friend above, so dr=
ops
in paper and
6. dials the number and presses start
So far this is why fax is so popular...look how simple! Why change thi=
s.
7. Fax machine at our friends house rings, answers and the two machine=
s
negotiate capabilities
8. Cuz fax standards are simple (speed is 9600 or 14400, image type is=
G3
or G4, resolution is fine
or super fine) the machines agree on a format.
9. The 'elsewhere' person's fax sends the image line by line, page by
page,
and
10. Our friend's machine receives and prints accordingly. When done w=
e
disconnect from the phone line at both ends.
<HL> 11. Someone gets to pay a big phone bill... later. ;-)<HL>
Now an Internet fax should act as much as similar as possible...except =
that
instead of phone numbers we
either use an IP address, or better yet a standard hostname (that DNS c=
an
resolve). The connection
between two machines could ideally work the same...i.e.
1. Internet fax at person A dials into ISP. Gets hooked up and "onlin=
e".
2. Internet fax A sends request for capabilities to internet fax B
3. Internet fax B is NOT online. Ooops. Now what
<HL> Woooa! You are beginning to loose me. Why did you assume the rcvr =
was
"there" in the "fax" example, but not "there" in the Ifax example!? Let=
's
talk apples to apples please. So, maybe your real question is... "what
makes me believe printers with URLs will be there?". Or maybe your issu=
e
is... "what if the receiver=92s life on the web is transient"? (Maybe
start by asking what if the fax machines life on the wire is transient)=
<HL>
4. Or Internet B is online (we happen to catch it while it is dialed u=
p),
so Internet fax B responds with
G3 or G4 then waits
<HL> Or PCL, or Postscript or PDF or... <HL>
5. Internet fax A now sends first page in TIFF wrapped G3 and waits fo=
r
response
or continues to send until the job is complete (I am not sure yet how t=
his
part works under IPP)
<HL> Probably uses chunked encoding to stream the entire job.
6. Internet Fax B receives and sends "Okay got it" as the response.
7. Machines "disconnect" from each other, but may or may not hang up t=
he
phone line.
<HL> What phone line? I think this is probably the root of our mismatch=
.
<HL>
You should see the key issue is the receiving fax machine (or any other=
Internet device ) that is not
full time on the network.
<HL> Or ANY device (i.e. conventional FAX). Same problem... right? How =
do
you solve it with FAX? Repeat that answer, in general terms, for this
problem.<HL>
a. What do we do if Internet device B is not online at the time
Internet
device A "calls".
b. Can a proxy of some sort (i.e. an IPP intermediate server comput=
er
(I
didn't want to write a server server))
help resolve this issue?
<HL> Sure - why not? <HL>
c. Do we want to instead require an IPP enabled Internet fax be onl=
ine
at all times (thereby being called a
full-time fax) or are there other solutions
<HL> Can we? Wait one while I run across the hall and unplug our FAX
machine.
Which door will the fax cops badge in at? <HL>
Where is Larry when you need him?
<HL> Probably groaning like many others who already understand the prob=
lem
and
know the answer. Again, I'm sure I will grow to respect the issues bett=
er -
and
it's obvious I'm striving for (perhaps over) simplification.
The way I would distill your main issue is... even when fax started, it=
was
a fair assumption that if someone gave you their fax number and you dia=
led
it
you could "send a fax". (I'd actually argue that this is not absolutely=
certain
even today... I've dialed fax numbers and heard dogs barking, dialed ph=
ones
and had fax spoken to me). At the beginning of IPP... you are wondering=
if
someone gives you their printer URL how sure you can be that, if you tr=
y,
you can send a print job.
This is a very good question... IPP won't work very well if the answer =
is
something like 20%, will it? So, the issue stands when we're discussing=
IPP, not just "fax" over IPP.
With 25 or so vendors demonstrating interoperability and product
development
efforts at the bakeoff a couple weeks ago, the issue is far from academ=
ic.
And with nearly every OS/NOS vendor represented in that number... I hav=
e
a feeling it won't go unaddressed! <HL>
Mike
Harry...
> -----Original Message-----
> From: harryl@us.ibm.com [SMTP:harryl@us.ibm.com]
> Sent: Thursday, March 25, 1999 1:29 PM
> To: Michael Crawford
> Cc: 'Richard Shockey'; ifx@pwg.org
> Subject: RE: Word docs
>
>
>
> I probably need to learn more respect for the problem but, right now,=
it
> sounds like far too many words than necessary.
>
> To send (legacy) FAX, I "dial in" and the fax machine has to "answer"=
.
> This
> is anologues to "logging on" (perhaps via dial to my ISP) and printin=
g to
> a
> URL. The printer has to "be there"... of course.
>
> In either case, if there's a wire laying of the floor at the receiver=
> end... no go.
>
> When I re-read your reply, I get the impression you are describing a =
IPP
> SERVER (printer) that is ISP dial connected. This would be different
> (still... not unlike some of these folks who use the same phone numbe=
r
for
> voice and fax. Sometime I call their "office" and get "carrier blast"=
,
> other times I get their voice mail... hard to know which it will be).=
>
> Guess I thought it was the following
>
> >IPP yes, QUALDOCs maybe not. I think you are right...IPP should be >=
like
> using a printer on a very long LAN (this doesn't suppose >10baseT by =
the
> way PPP dialup should qualify!).
>
>
> >Of course it is assumed that any IPP 1.0/1.1 client could use a >dia=
l up
> connection.
>
>
> that got us on this thread and this is obviously referring to a dial =
in
> client.
>
> Harry Lewis
> IBM Printing Systems
> harryl@us.ibm.com
>
>
>
> Michael Crawford <mcrawford@iready.com> on 03/25/99 02:10:59 PM
>
> To: Harry Lewis/Boulder/IBM, Michael Crawford <mcrawford@iready.com=
>
> cc: "'Richard Shockey'" <rshockey@ix.netcom.com>, Michael Crawford
> <mcrawford@iready.com>, ifx@pwg.org
> Subject: RE: Word docs
>
>
>
>
>
>
>
> > -----Original Message-----
> > From: harryl@us.ibm.com [SMTP:harryl@us.ibm.com]
> > Sent: Thursday, March 25, 1999 12:35 PM
> > To: Michael Crawford
> > Cc: 'Richard Shockey'; Michael Crawford; ifx@pwg.org
> > Subject: RE: Word docs
> >
> >
> >
> > REgarding...
> >
> > >*** Hmm. Isn't it difficult to have a "server" application that is=
n'
> > >there to serve (i.e. not dialed up when you request service)?
> >
> > We were talking about the CLIENT dialing in (I thought). Not the IP=
P
> > server.
> >
> Yes, but in IPP it takes two to tango...the client must find the=
> destination
> up in order to exchange capabilitiees and to send data...there i=
s no
> notion
> of store and forward, at least not in the present form of the sp=
ec.
>
> > It would be equally
> > difficult to send FAX to a machine that was unplugged to from the p=
hone
> > line... no?
> >
> Yes, but a legacy fax is usually plugged in BUT NOT ONLINE...i.e=
. it
> answers.
> We need to figure out how we answer a request for service by the=
> target fax machine
> or specifically say in the specification how a fax machine MUST =
be
> online in order to
> provide services. The implication is that:
>
> 1. An ISP must dial out (Don't hold your breath)
> 2. A IPP enabled fax periodically comes online (dials up) to=
> receive messages creating
> short window for IPP connection to occur
> 3. An IPP enabled fax MUST be on the LAN (oops, this means
> 10BaseT connectivity not dialup) which
> is in turn reachable by the source machine (WAN or Intern=
et
> gateway with hole in firewall).
> 4. A fax service will act as the destination and then forwar=
d to
> the true destination when it periodically
> comes online by dialing up.
>
> I like LAN connected faxes, but there is that cost issue once ag=
ain.
> Higher end Internet fax solutions
> will undoubtedly be on the LAN (look at the first internet fax
> offerings today...LAN connected, expensive
> and volume which hasn't yet paid for the datasheets let alone th=
e
> development of the machine).
>
> Mike
> > Harry Lewis
> > IBM Printing Systems
> > harryl@us.ibm.com
> >
>
>
=
--0__=oY3Zm2toYZe8fqdRVkuGGdAEz1QoQ2pC8hsoIPkkdnFl9fBBoaDpfjpI--