Subject: Re: PDF limitations From: Rex Ballard Date: Mon, 26 Sep 1994 10:52:45 -0400 (EDT)
How the Web Was Won
Subject: Re: PDF limitations From: Rex Ballard Date: Mon, 26 Sep 1994 10:52:45 -0400 (EDT)
To: ace@tidbits.com
Cc: online-news@marketplace.com
Mime-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: jvncnet!marketplace.com!owner-online-news
Content-Length: 14307
Status: RO
X-Status: 
X-Lines: 280


On Fri, 23 Sep 1994 jvncnet!tidbits.com!ace@dowv wrote:
> >Well, that raised the hairs on my neck. (I'm a former newspaper graphics
> >editor, so you pushed a button. 8-)
> Yeah, thought I'd get a response to that. :-)
> 
> >I *might* buy that argument when applied to the Wall Street Journal. The
> >quality of its reporting outweighs any discomfort I have in its (IMO) ugly
> >format. In most any other publication, if it *looks* like it was put

Interesting point here.  The Journal (Highlights) are available in
electronic media form.  There are a variety of different presentations
available, depending on other desired services and target systems.

> >together by high school students, I start doubting the quality and
> >veracity of the text. If the design is really awful, and inhibits my

There is a good bit of "Look and Feel" that is very flexible.  To an
investor, receiving real-time news feeds, the presentation is almost
secondary.  In fact, any beautification that delays makeing the story
readable as quickly as possible is considered a liability, not an asset.

> >ability to find the information I want, I'll probably find a competing

Dow Jones stories are reviewed by editors and tagged with unique search
keys which are dilivered in a separate field from the rest of the story. 
In some products, the user never actually sees these keys, but the user
can select specific industries, companies, and/or regions to make the
article easier to find.

> >publication to read. Like it or not, the design of your publication (if

When stories are breaking at a rate of 5 to 10 stories per minute, it is
important to make a small part of it readable (usually just the headline).
 Even after you've narrowed the search, a flurry of stories affecting a
specific industry or company can quickly swamp your "reader".

> >it's bad) will subtly turn away readers; if it's good design, they won't
> >even notice it and your pub will give the appearance of being produced

What is most important is to identify the particles of the story
(service, publication, headline, leading paragraph, story paragraphs,
story tables, catagory codes...).  This is what must actually be shipped
across the wire.  Refer to ANPA standards as an example.

Everything else is a matter of taste, and intangible. 
The V.P. of marketing might want the article formatted for Lotus Notes
with a Newspaper Logo, fonts, and key search words highlighted with color
or fonts.  The investor, or marketing person responsible for tracking
activities of compettitors may want minimal formatting.

Identifying the objects themselves provides a good way of managing the
level of detail.  Even such simple issues as pictures can be handled by
giving the HTML reference, a small "iconified" picture, and the full 1024
by 768 256 color graphic.  It's ironic that the quality of such a
presentation is actually so good that it can't be put into hard-copy.

> >by professionals. Yes, the text is paramount. Is design irrelevant? Of
> >course not. Good design supports your precious words; bad design (and by
> >that I would include *lack* of design) undermines them.

More important than "Layout and Fonts" are issues such as editorial
guidelines.  Making sure that attribution, by-lines, and headlines are
properly handled.  Making sure that tables don't get rejustified into
paragraphs.  Providing date, sequence, and time of the article so that it
can be requested later from a database, and other related forms of
information are critical just to cover the legal and service requirements.

> You're of course right about the importance of design in paper
> publications, but look at the huge number of people who read Usenet news
> (or even the 65,000 or so that read ClariNet), where there's essentially no
> design whatsoever.

With usenet news, you get what you pay for.  You get about 4 gigabytes of
flames and about 20 megabytes of reliable information you can actually use
as the basis of making sound career, professional, or business decisions. 
It is good to know what's churning in the rumor mill, but I wouldn't want
to bet the next 10 years of my professional life on a rumor that Sun might
port Solaris to the Alpha chip.

What is fun is to request verification from an official source on some
article or posting to usenet.  Even my postings to this list are not a
reflection of policies or practices of Dow Jones (my employer), any
similarity is merely a coincedence.

For anything significant to be accomplished, there are 5 types of
conversation that must be conducted:

	1.  Conversation for relationship.  Who am I? Who are you?  What
	    interests, goals, values do we have in common?  On the
	    internet, we have alt, talk, and soc groups that specialize in
	    this type of conversation.  Ironically, even the flame is part
	    this type of action.

	2.  Conversation for possibility.  This is when we begin to
	    what might be possible for the future.  How can we fulfill
	    common goals.  Even "problems" are actually possibilities
	    expressed as limitations of the past.  This is where advertizing
	    has traditionally occurred.  A very effective writer can send
	    a reply to a heated discussion which shifts the converstion from
	    a gripe session to an almost endless list of opportunities.

	3.  Converstion for opportunity.  This is when the possibilities are
	    narrowed to several specific opportunities which may require
	    some sort of effort or action.  Real-time news feeds are a
	    nearly endless stream of opportunities.  They invite the reader
	    to take action (buy or sell stock or products, give loans,
	    contact competitors, or change/start product plans.  To a
	    real-time reader, about 1 out of every 5 articles is an opening
	    for action.  Usenet groups rarely become an opening for action.

	4.  Converstion for action.  When an opportunity is considered worthy
	    of action.  The actions to be taken consist primarily of making
	    requests and promises.

	    In the 60's a purchase order might be sent
	    with the implied promise of payment upon receipt of the goods.
	    A confirmation of the order would be received.  In 6 to 8 weeks,
	    the merchandize would be received.

	    In the 80's the order would be faxed and the payment issued by
	    credit card and  merchandize shipped Federal Express in 2 or 3
	    days.

	    In the 90's we can send the PGP encrypted EDI transaction via
	    computer and their computer will request the pickup and process 
	    the meter of the package for shipment.  The merchandize will be
	    there by 10:35 the next day.

	    Some merchandize, such as information, periodicals, stock
	    transactions, and software can be delivered within 3 seconds of
	    receiving the order.

	A simple practice such as adding an HTML reference to an "advertizer's"
	catalog can make this type of instant purchase power possible.
	The challenge will be that the user can literally choose from nearly
	anyone in the world capable of producing that product.  There will
	probably be an opening for "Supercatalogues", which are just
	references to local products in 20 or 30 different localities.


> But I'd also argue that there are some simple design considerations that
> work just fine in ASCII text, things like plenty of white space and
> relatively short paragraphs, so I'm not actually saying there should be no
> design, merely that things like Acrobat aren't necessary for what design
> should be done.

Not only are they not necessary, they actually get in the way.  What is
important in a publication is that you know source, product, headline,
story, tables, pictures, and graphs.  What gets in the way is trying to
determine how to display 18 point Roman Serif on a vt100 emulator.  Worse
yet, being stuck with only one possible display client, that takes 15
seconds to format stories that are coming in every 5 seconds.

> In TidBITS, we use the setext format, which is 7-bit ASCII
> and quite flexible, but still builds in some design considerations (nothing
> longer than 65 character lines, two-space indents before paragraphs of
> wrapping text to increase whitespace, and so on.)

Other examples of this are nroff (unix), script (mvs), and TeX (gnu). 
While these are great formatters, they don't tell you what type of object
you are dealing with, how to deal with changes to a story (retractions and
corrections), or the difference between the leading paragraph and the
second paragraph.

> Although there are
> probably people who don't read TidBITS because they find it ugly, I've
> never had a comment to that effect. In fact, that wasn't true way back when
> we used HyperCard to control the fonts and display - people hated having to
> read in the fonts I had chosen (even though they were the best for reading
> on the screen in most opinions) and so on...

One would think, after all the furor of customization of look and feel
that was demanded in X11R4, Windows 3.0, OS2 2.0, and Windows NT, that it
should be obvious that user's will almost always want to customize any
display they read regularly to support their cards, display sizes, and
personal preferences (I might like 6pt helvetica, but I'm not nearsighted and
I have a large display, try that with a 50 year old man and a 14 inch 8514
interlaced display and he will not like it).

The market for WYSIWYG News is corporate executives who will tell the
proof-reader that they want their news to come in their favorite fonts. 
It is a perfect career killer to say "I'm sorry sir but the vendor does
not support that".  It is important to have the ability to switch vendors,
access GPL source code, or modify all configuration files.

Open Systems are defined by published and freely circulated standards
supported by freely circulated source code.  Often, vendors will add
"Glitz" to their products.  The best example is the way Microsoft Access
can make a basic SQL server look like a high powered GUI application.


Standards such as ANPA, HTML, SGML, and MHEG are generic enough that the
base objects can be identified.  These also include GPL readers that can
be used to format the objecs into "pretty" formats.  Some vendors do require
a basic non-disclosure agreement before you receive the free source code
and specifications.


> As an aside, ancient Greek was actually written in all caps with no spaces
> between the words. Talk about a design nightmare.

Ancient Hebrew was written with no Vowels.  You couldn't even be sure that
a word (taken out of context) was being pronounced properly.

> >The online publishing business is so new, few competent designers seem
> >to have arrived on the scene yet.

There are actually 30 or 40 companies who design on-line publications
using the same feeds.  I am always fascinated by the diversity with which
each vendor can use this information, even within relatively strict
guidelines.

Motif is currently one of the few really widely available multimedia news
"readers".  Often proprietary feed formats are converted to HTML for display.
The result is a very attractive display with minimum feed overhead.

> > I wouldn't be so quick to dismiss what
> >these people can do for your online publication. Take a look at some of
> >the Web pages currently being done by professional publishers. Not very
> >inviting.
> Pretty amazing how awful most of them are, isn't it?

This is relatively new capability.  Remember, the Journal still only does
those "dot pictures".  Most newspapers are not accustomed the the concept
that they can put 256 color graphics into their publication.  As the
competition heats up, we will see more "magazine quality" publications.

The important thing is that there is no mandate on how the information
will be processed.  Unlike a printed page where each article can be laid
out as a complete page, the primary attraction of the electronic media is
that the reader can filter the publication so as to create a "custom
newspaper" targeted to his interests and needs.  Many readers have several
such "profiles".

> And just to tie in some threads, here, consider how much time is spent
> creating the information for which people are actually reading the
> publication versus the amount of time spent formatting and tweaking and
> mucking with copy to make it fit around the graphics correctly.

Remember, in electronic publishing, there are two kinds of publications. 
The real-time publication or "wire" is often a race in real-time.  Often
the editor has 30 to 90 seconds to code the story for effective filtering.

In the "archive" publication, you are dealing with dailies and weeklies that
can spend more time on lay-out of each article.  On the other hand, you
are "dressing up" the best of about 1% of the total flow of information
coming in.  What gets "dressed up" is what the editor, in his subjective
opinion, considers most valuable and newsworthy to his readers.

An archive database client however, searches the "dressed up" stuff, along
with the other 99% of the information base.  The ultimate result is that
you will rarely see only the "pretty" stories.  Most will be "Information
in the raw".

> I don't
> know what the ratios are, but I do know that if you minimize the time to
> lay out the publication, you've got that much more time to make the
> information good before the deadline. Most of my formatting and layout is
> done via macro, and takes about three minutes. If I had to spend several
> hours flowing into PageMaker and tweaking, that would be a significant
> amount more work.

The trick here is to be able to say - "Display the bitmap for each product
logo, and each source logo, display the headline in a big dark font, and
the leading paragraph with the first word in a double size font.  Now,
another vendor may just say - display the source and product on one line in
bold, display the headline on the next line in bold.  Display the leading
paragraph in roman, display the rest of the story in italic.

The formatting can be simple - ala SMTP headers, or complex ala MHEG and
SGML definitions, or even composite feed.  What shows up on the user's
screen can be a nice pretty acrobat display of his choosing.

> cheers ... -Adam
> Adam C. Engst, TidBITS Editor -- ace@tidbits.com -- info@tidbits.com
> 




From jvncnet!marketplace.com!owner-online-news Tue Sep 27 16:36:08 1994