SUMMARY (very long): Preferred Online Formats

[Sandra Charker <scharker -at- connectives -dot- com>]

In January I asked about TechWhirlers' experiences as users of PDF 
documents online. The specific questions were:

1. Do you notice a difference between reading PDF online and reading HTML 
online?
2. Do you have a preference either way? If so, which one?

Here's the summary or responses and a synopsis of further information I've 
been tracking down. I'm sorry it's taken so long: like many an unplanned 
project it got way out of hand. I've tried to restrict this message to 
issues of interest to the list; if I've failed please accept my apologies.


Summary of responses:

12 people responded. All either preferred HTML themselves for online 
reading, or assumed that most people do and suggested ways to improve PDF 
for online. Several like Acrobat's search capability. Some mentioned that 
printing a PDF gives a much better result than printing from a browser.


Attention-grabbing sound bite:

For people reading online, PDF has no advantages and some substantial 
disadvantages over HTML. The benefits of PDF are all to designers, 
publishers, authors, and content controllers. For designers those 
advantages are overstated: at this time there is no such thing as control 
over the appearance of an online document. For publishers, authors, and 
content controllers, the advantages can be outweighed by user difficulties 
with readability, file download, and the Acrobat reader.  There's good 
evidence that people read, comprehend, and retain information less 
effectively from screen than they do from print. PDF compounds this 
problem. At this time PDF should not be used as an online format except for 
ephemeral, archive, or trivial material.


Balanced and moderate discussion:

I have no argument with the value of PDF as a format for documents intended 
for print, and particularly as a means of distributing print-ready copy. My 
questions and this discussion are wholly concerned with reading and using 
PDFs online.

However, I question the assumption that general users print PDFs. I think 
it's probable that many are read or skimmed online before their readers 
make a decision about printing them.  Given the popularity of the Acrobat 
search, it's even possible that *most of the PDFs published as printable 
documentation are used mainly online.

In digging around for information about the effective use of PDF online I 
found myself exploring the world of e-books (devices and content), yet 
another area where technology is moving with breathless speed and the 
ongoing war between proprietary and open standards is raging. Two points 
are relevant.

First, PDF is a widely-used format for commercial electronic books. 
Publishers sell PDF books with the explicit expectation that they'll be 
read online. Writers' e-book websites recommend PDF because the pages look 
like print.

Second, there's an immense amount of effort going into improving resolution 
for online reading. We discussed e-book devices and the ClearType format 
briefly on this list late last year. I've recently seen a report of a 450 
dpi device; blurriness probably isn't a problem on that one. OTOH, Adobe 
and Palm have just announced that the PalmOS will support PDF; that's 
mind-boggling, even to someone who's happily read magazines and books on a 
Palm III.

The problems with PDF seem to fall into two groups, PDF readability and 
Acrobat Reader usability. There's a also a set of issues to do with the 
relationship between Web browsers and Acrobat.


PDF Readability

No one liked PDF better than HTML for online reading; some had no 
preference, or their preference depended on what they wanted to do with the 
text. Comments included "it's painful to look at", [PDF files] "are 
typically hard to read on-screen, particularly if you have an older 
notebook computer with a small, poorly-lit screen", and "the blurriness 
causes online PDF to lack the 'immediacy' of HTML characters online".

I haven't found any discussions of PDF readability on publishing or 
usability sites. This surprises me, because it's not hard to find people 
around the Web who don't like it. I haven't kept all these links, but 
common words are "blurry", and "smeary". The following is a slightly edited 
msg from a TW who works for a software house and is a long-time advocate of 
Acrobat. I quote it because it's the closest thing to quantitative 
information I have:

"... A third to half the people [at this company] using Acrobat find it's 
not optimum for long term screen viewing due to the anti-aliasing. 
Personally, I don't even see it... Interestingly, it's women who seem to 
have the most complaints. I think I found one woman (out of the eight that 
work here) that didn't mind Acrobat's display. All the others found it a 
little "soft" or "fuzzy." Of the 5-6 men I've asked, a couple have 
complained of resolution but most haven't. The biggest complaint is that 
they don't like the software--they prefer WinHelp."

Common suggestions for improving online reading include setting the page 
size and shape so that it matches a monitor better than an 8X11 or an A4 
page, and using a larger than normal (for print) font size. These things 
undoubtedly go part of the way, but even for page size and shape they 
aren't the whole story: what's an appropriate page size and shape for a 
mobile phone? For a Win-CE handheld? And for fonts,  "Use larger fonts" 
seems to be about as complete a guideline for usable PDFs as "Use active 
voice"  is a complete guideline for clear writing.

Here's a test. PDF some text in one of the fonts specifically designed for 
the Web, and compare the PDF output with the same text in HTML. I used 
Distiller to PDF text in Georgia: for the PDF I used 10, 12, and 14 
point.  14 pt is about the smallest recommended for an online PDF, but I 
tested the smaller sizes because even 14 pt seriously limits the amount of 
text that will fit into a display. Didn't matter: even at 14pt, Georgia, 
which is my preferred screen font, makes me feel slightly queasy in PDF.

This highlights one source of my confusion about this whole issue. I have 
reams of information about fonts for screen and fonts for print; I can find 
many discussions about font processing for postscript. But I haven't found 
any information about fonts that will be processed through postscript to 
screen. And, as the Georgia test shows, that process can break the rules of 
font selection for online viewing.

I therefore disagree with the claim made by one respondent that "A PDF 
designed to be read on screen can be every bit as good as HTML, and 
probably better (on the assumption it's designed by a typographer who 
understands that legibility takes [precedence] over creativity)". I think 
that the technology is not yet up to the requirements of typography.

I did find references to some font traps for producers of PDF files. One is 
to do with Type 3 PostScript, which apparently can lead to problems when 
the output is viewed under Windows. Another is to do with documents 
originating in LaTex. In both cases the output is described as blurry; I 
don't know how they compare with the Windows-produced Windows-viewed 
blurriness that I experience.

Adobe also has a couple of support database articles on blurry output. 
Their solution for blurry text is for the user to turn off anti-aliasing. 
I've tried that - not nice.  More to the point, a general user who doesn't 
know PDF from pickled dill is unlikely to find this "solution".


Acrobat Reader

The person who needed to prevent users from printing their online documents 
was one of those who mentioned that many people "don't know much about 
using the Acrobat interface & aren't interested in learning." Speaking for 
themselves, others commented that they don't like the scrolling in Acrobat, 
that they don't like the zoom options, that it feels clunky.

Many people do like the ability to search multiple information chunks (IOW 
a book, if that's what the document is) instead being limited to searching 
the current page as you are in an HTML browser.  This fits  the experience 
I had on a recent contract: developers regularly used the printer-ready 
PDFs of user manuals, even while they complained about how much they didn't 
like them, because they like the search. That example was interesting 
because the identical content was also available to them as HTMLHelp, but 
they didn't seem to like that much either - I don't know why.

Incidentally, you can't run a full-text search on a printed document. It 
would be reasonable for people who like Acrobat for its search capability 
to prefer to use PDFs online.


The Reader and the Browser

By definition, if you launch Acrobat either in a browser or from a Web 
page, you're breaking away from the conventions of the Web. That's risky: 
the Web is confusing enough already without introducing a whole different 
look and feel. If, on the other hand, you provide PDFs for download, you 
rely on users to have downloaded and installed the Reader for themselves. 
That's also risky: it's a big download, and PDFs are often big downloads 
too (though an awful lot of sites don't bother to tell you how big their 
files are).

However, my impression is that most TechWhirlers are using PDFs either for 
intranet material or to supply as user-printable manuals with software.

For an intranet, I think you still need to be careful, especially if the 
organisation is widely dispersed - and who has more need for an intranet 
after all? Connection speeds vary, technical support varies (and yes there 
are people who need help installing Acrobat reader), and equipment varies - 
even sites that standardise on software and versions still have old machines.

For both intranets and with software, do not assume that users know how to 
use Acrobat, and don't assume that they can or will print your documents, 
even if you tell them to.  As far as possible, design your documents so 
that they are at least legible online - the best solution for reasonably 
sophisticated users is probably 2 versions, but at the very least you can 
avoid fonts with fine strokes.  And if you get the chance to do usability 
testing or studies, or even surveys, of how your users really do use your 
PDFs, please grab it. I haven't found any quantitative information about 
PDFs: if it exists at all, then I think it might be proprietary or 
commissioned research.


Where to now?

One TechWhirler commented that "Acrobat's PDFs are a stop-gap solution that 
will happily go the way of DOS and GOPHER." I agree that PDF is interim and 
very imperfect technology, and I think that the paper-centric model of PDF 
is fundamentally flawed for online use.

Which was once all fair enough: PDF preceded the Web, and had wide 
acceptance in some major organisations (the US government for instance) 
before the Web went mainstream.  It was reasonable then to expect that 
thousands of people would become familiar and comfortable with Acrobat 
Reader; nobody knew in 1992 that millions would become familiar with Web 
browsers.

PDFs are quick to produce from widely-used authoring software, can preserve 
scanned material, and do cross platforms. Those are real benefits. And, 
while I think the design benefit is overstated, I don't think it should be 
ignored. Book design, like book indexing and cataloguing, has developed 
over hundreds of years. There's precious little evidence yet that online 
design has taken that knowledge to heart, let alone started to move forward 
into its own future.


Thanks to TechWhirlers:

Jessica Lange
Tim Altom
Blaine Bachman
Darren Barefoot
Joy Brady
Daniel Hall
Geoff Hart
Paula Jagt
John Locke
Chuck Murray
Salatas
Jean Weber

and special thanks to Jerilynne Sanders of Simply Written


References and Resources

The following books all include some information about typography or font 
selection. Some of it is pretty cursory, and none refers specifically to 
PDFs. OTOH, they're all well worth reading for the rest of their content.

Karen Schriver : "Dynamics in Document Design"
Marlana Coe:  "Human Factors for Technical Communicators"
William Horton: "Designing and Writing Online Documentation"
Joann Hackos and Dawn Stevens: Standards for Online Communication"

The only book I've consulted that deals specifically with typography is
Robin Williams: "The Non-Designer's Design Book"

I've just ordered a copy of:
Alex White: "Type in Use: Effective Typography for Electronic Publishing"


Typography:
http://www.graphic-design.com/Type/ Print-focused, but most typography 
sites are.

http://www.will-harris.com/typoscrn.htm This page is Daniel Will-Harris' 
discussion of type for online. The site includes a clever guide to font 
selection and a heap of material about fonts in general.

http://webreview.com/wr/pub/Web_Fonts Daniel Will-Harris features heavily 
on this site too, but the emphasis is on type for the Web.

http://www.microsoft.com/typography/ Not a lot here, but there is 
discussion of Microsoft's ClearType implementation and a bit about 
Microsoft's take on the Open eBook standard.


Sites about PDF:

Fascinating to see a tide of interest in PDF for online forms... Look out 
guys: the inmates are targetting this asylum as well.

http://www.pdfplanet.com/ A world (sorry) of expertise. The subsite 
http://www.acrobuddies.com/ is where I'd ask for help if I had an online 
project where there was a serious problem with HTML.

The e-paper section of planetpdf  contains a link to a PDFed white paper 
"Using Large Format PDF", which is about PDF for online. It uses a 20-point 
sans-serif for the body text, and to my eyes most of the vertical strokes 
are blurry. The text is certainly readable, but the price is that not much 
fits on a page.

http://www.performancegraphics.com/ This is a company site, with a lot of 
detailed technical information about PDF itself and effective ways to use 
it. It's a bit confusing, and has a painfully large graphic on the home 
page, but it's worth pursuing.

http://www.pdfzone.com/ Is a zone bigger than a planet? Dunno, but this one 
also has a lot of good info.

http://www.adobe.com/support/techdocs/3ece.htm Adobe itself has plenty of 
information about PDF. This particular link is to the article about blurry 
text.


eBooks:

http://www.openebook.org/ Open eBook is the non-proprietary standard. It's 
based on XML, and its explicit intention is that any eBook can be read on 
any device. The specification is available for download in *seven formats: 
subtle aren't they? The DTDs are also available.

http://www.jeffkirvin.com/writingonyourpalm/ An eBook author with 
interesting things to say about formats and a useful set of links to eBook 
sites. The site also has some really useful ideas for using the Palm for 
writing.

http://www.ebooknet.com/ A lot of information from a business perspective 
about e-book developments.


Web usability:

Jakob Neilsen: http://www.useit.com/
Keith Instone: http://www.usableweb.com/
David Anderson: http://www.uidesign.net/ A different perspective on some of 
the issues involved in content delivery.


Hope all this is useful to someone. Thanks again.

Sandra Charker

mailto: sandra -at- connectives -dot- com
http://www.connectives.com/