Jont Allen's comments on the merits of different formats for electronic conference proceedings

I am the chairman of a small conference, and we have been looking at the question of how make a CD-ROM of our proceedings. It is clear that many others have been faced with this challenge lately, and we have all witnessed the results of their efforts. Most of these CD-ROMs have adopted AcroRead as a reader. I am not totally clear why this has become such a popular solution. My analysis is given below. I am interested in your thoughtful opinions.

Review of my assumptions:

Here are some reasonable alternative CD-ROM formats that we are looking at:


I have yet to see what I would call a great CD-ROM. The real issue here is good human-factors (Ergonomics). In my view, the experience should be the best of the WEB. This is where new browsing technology has been most heavily tested.

The issue is not about format (i.e., pdf vs postscript). While one could make an issue of this, I will not be sucked into this black hole, at least not here and now.

There are three basic experiences one has in using a CD-ROM:

All three are important. If you can't easily read the material, there is no point in having the disk. If you cant find what is on the disk, it does you no good. Printing must be painless.

Reader-software is important because the human reader must maintain his/her concentration as well as see the material from the screen without eye strain. Browsing is important to find things efficiently and quickly. Printing is important because we always will want paper copies when we are away from our computer.

I doubt it is possible to agree on a rank order of most important, and why bother. Lets say, for sake of argument, the three are equally important.

The Reading Experience

Some alternatives are:

The Browsing Experience

Netscape's hypertext browsing technology is a clear winner. Netscape has made this point so effectively that Microsoft has adopted a related technology as the basis for Windows 98 (99+?).

Acrobat allows hypertext. However is inflexible since you can't easily change it without a special purpose expensive tool you must learn. The Acrobat solution uses a proprietary binary file format. HTML editors are common, and are based on an open (nonproprietary) ASCII hyper-text.

Nobody would consider AcroRead as a serious web browser.

The Printing Experience

I see this as a potentially complex issues that I have not properly dealt with here because I need to study it more.

There are two basic problems, format input and format output. Format input is always postscript for both Acrobat (via distiller) and ghostview. PDF is an intermediate format, based on (modest) extensions of postscript.

Output format depends on the printer (i.e., postscript, hpgl, epson, etc.)

The print-server manages the print spooling, and some formating issues. It is OS dependent in all cases I know of.

I don't understand how acrobat can support all types of output formats. Does Acrobat act as a print server? Does it depend on the OS to support the printers? If so, then it is as weak as the OS on this issue. Or does it have all the printers that exist internal to the software? In this case what happens when a new printer comes out (i.e., a new 1200 dpi color printer for example).

Postscript and the WINTEL platform

My feeling is that the postscript printing problem remains largely unsolved on the WINTEL platform. This is an OS issue, not a postscript issue.

For example, without GSview, printing postscript is unnecessarly difficult on a WINTEL box. Since postscript is an ASCII file, one should be able to use the DOS print command, or COPY from the file to LPT:
Win95 claims that LPT1 is write protected. Can anybody tell me how to solve this problem?

I believe that Microsoft has systematically done everything it legally can to discourage the use of postscript on the WINTEL platform. For example the Microsoft postscript does not conform to open and well documented DSC long-standing conventions (it is not conforming encapsulated postscript), and they have allowed it to be confusing to use, slow, and memory consuming. For example, my wife generated a one page postscript memo in MS-Word from Office that took over 3 MB, and overflowed the printer memory. It only had a few (i.e., 10) words in it! Ghostview viewed it without problems, but we couldn't print it because we didn't have enough memory in the printer. She, and then I, wasted more than an hour on this trying to print the one page memo. (It problem turned out to be the bitmap font used by the memo wizard, not the postscript printer, which is wonderful.)

Unix, coupled to a postscript printer, is seamless at this point. (I set up my own printcap file with only a few minutes of work, and it worked the first time, and has never stopped working. I'm not saying that it would be that easy for everybody.)


Browsing a CD-ROM with Netscape is a wonderful experience, and is platform independent.

For many people, and I don't know how many, Ghostview/GSview is an excellent screen-rendering solution. It gives the flexibility of postscript at a low cost (in either time, money, or disk space). The problem has been setting it up on the WINTEL platform. (This probem has been solved, but the solution is relatively unadvertised. Check out the proceedure given my home page.)

The frequent claim that postscript is slow is uninformed and misleading. These claims are probably based on WINTEL problems, a slow serial interface, or an old printer having a postscript cartridge. (Uncompressed postscript files are typically very large.)

In conclusion

My bias:

Acrobat does not get high scores from me either as a browser or as a reader. I don't know much about Acrobat as a printer filter. Ghostview/GSview is a screen rendering program, not a browser. It is an excellent printer filter that supports virtually every useable printer made.

What do you think? Please send me your response.

Jont Allen (

Back to Mohonk WASPAA'97 hints & tips