In-Reply-To: 
Message-ID: 
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Status: O
X-Status: 



On Sun, 30 Apr 1995, Cliff Allen wrote:

> R Ballard  writes:
> 
> > 
> > Can you tell me where I can get either:
> >         A.  Precise specification of PDF - comparable to the PostScript Book?
> 
> Adobe talks about the specification be "open" but I've not seen the spec on
> their Web site or their very detailed CD-ROM demo.  I suspect it'll take a
> call to Adobe asking for the product manager for Acrobat.
> 
> They provide very good specs on the API for Acrobat Exchange in order to
> create add-in programs in Windows and Mac environments, so it is possible to
> create custom applications that use PDF without having to write the PDF files
> from your programs.  (I'll be giving the API specs to our UNIX programmer for
> him to review for possible PDF development here -- but he's not up on this
> topic yet.)  They also have other software development kit (SDK) products for
> Acrobat -- I just can't remember what they are.

APIs are great when you are an application programmer who just wants to 
"convert the postscript to PDF bits and ship them".  When you are 
responsible for the integrety of the transmission of 16,000 PDF 
documents, and 5% of them are unreadable, you must be able to understand 
the content.  If they are compressed and encrypted, give me the source to 
decrypt and decrompress them.

I've spent from 9:00 PM to 3:00 AM 5 weekends in a row baby-sitting 
transmissions of PDF files that "kill" part of our distribution chain.
(And you've been wondering why I could put out so much e-mail :-)

> >         B.  PDF viewers for Mosaic on Ultrix (4.1-4.3), Linux, AND all
> >             versions of Windows (3.0 to 9? and NT) AND OS/2, for modest fees.
> 
> The beta software WebLink allows a Macintosh user to connect Acrobat Exchange
> (a PDF reader/editor program in the Acrobat line) to a Web browser, such
> as Mosaic, Netscape, etc.  They say WebLink isn't available for Windows for
> some technical reason I don't remember.  They didn't even mention any of the
> UNIX versions of WebLink on their Web site.

That's right.  We have several customers who were thrilled to get a 
product of ours, only to find out that they couldn't convert the PDF to a 
more "usable" format because exchange wouldn't allow it.

> Netscape has agreed to incorporate PDF support, so I've got to believe it'll
> be in versions for all three platforms.
NetScape has started to support all kinds of proprietary platforms and 
content.  If they keep it up and drive "Open Systems" out of their market,
those who created the infrastructue will just use the source that is 
available and enhance THAT.  IBM lost the SNA market because they only 
supported 3270 and CICS protocols.  With TCP/IP they could send anything 
to any brand of host.

> >         D.  A SGML -> PDF generator that can work on VMS, OSF, or Unix
> >             and produce 3000 pages (charts, tables, fonts...)/day.  This is
> >             urgent, 3 NTs chugging away is producing about 300/week.
> The new scanning-to-HTML program in the Acrobat line (can't remember the name)
> might do this...but it assumes you have printed documents to OCR.  Probably
> not a good fit.
Not a bit.  We are going with SGML because it gives us control over the 
format and portability (direct to HTML and Postscript).  Unfortunately, the
PostScript to PDF is not available for say - a DEC Alpha or an HP-9000 
running UNIX.

> Since there is probably an SGML reader on the Mac or Windows, I'll assume you
> can tell the SGML reader to "print" to a PDF file using the Acrobat printer
> driver.  The first step is find an SGML reader for those platforms.

The problem is I have LOTS of information that needs to be chugged out 
reliably.

> I am unfamiliar with how to create a PDF file on a UNIX platform.  If the
> process uses a printer driver, then the same process could be used.

You can't create a PDF file on a UNIX platform!  They don't even support
SUN (80% of the financial services market).

> >         G.  An "acrobat accelerator" that lets me page through at 10
> >             pages/minute or faster.
> Not sure what you're looking for in an "accelerator".  You may be wanting to
> mark text segments as parts of a story so you can use the "follow story"
> feature.
I'm dealing with Stock Brokers and Financial Analysts.  They are trying 
to BEAT THE CLOCK.  They analysts has 15 minutes to respond to a market 
fluctuation.  If he has to wait 30 minutes to look at the 5 related 
documents which will guide his final reccomendation, our product is 
worthless (we'll just have to ship them plain text and JPEGs).

> >         H.  A "PDF to FAX" converter.
> I would think that a Acrobat Reader could be set to print to a fax/modem
> driver so you could fax PDF documents, but I've only faxed from Word, etc.
Unfortunatly, this means you must ship the fax to another machine to 
receive it as FAX (remember MS-Windows doesn't do muli-port telecomms well).

The customer that wants to do this has to process 5000 detailed 
documents/day and distribute them to several hundred clients as search 
responses.

> >         I.  A "PDF to GIF/JPEG" converter.
> 
> This one cold be difficult.  We create GIF files from the original art
> documents (e.g., QuarkXPress, etc.) as a separate task from creating PDF files
> from the art.  This might work for you.

Of course it would, if we were the tranformation and distributor.  The
issue is -- do we ship PDF, Postscript, SGML, or HTML to the customer?

> Since Acrobat Exchange allows you to copy both text and graphics from a PDF
> file (if not locked down with security), then I'd assume a graphic could be
> pasted into a program that saves in GIF/JPEG format.  We've never wanted to go
> in that direction, but I think it's possible.  I'll have to bring up Exchange
> and Photoshop and try this.

We've tried several approaches.  We want something that requires no 
manual intervention (we're processing 15,000 documents/week, remember).

> >         J.  A "PDF Validator" (Am I sending out garbage or is my link 
> > fuzzy)?
> Good questions, but I don't know how this would be done from within the
> Acrobat environment.  If the PDF file were put through PKZip (PC) or StuffIt
> (Mac), then the CRC checking of those programs would identify any file
> problems of transmission.

That doen't tell me if the data is valid in the first place.

> If you have any applications for Acrobat on the Internet, or feature
> enhancements for Acrobat to be used on the Internet, that are not
> confidential, I'd appreciate your comments.  We are preparing a document for

I can't say much (Nondisclosures), but I can say that we will ultimately 
have to manage a database of over 35,000 PDF documents which are updated 
at a rate of about 1000/day.  These then have to be shipped electronicly 
to several hundred users connected through several distributors.

> Adobe's Acrobat people on things for them to look at with regard to electronic
> distribution and use of PDF documents.
Handling of Large Volumes is a BIG issue.


> Cliff

Great Web Page!

	Rex Ballard


From rballard@cnj.digex.net Mon May  8 01:36:39 1995