My experience is that one can obtain much superior PDF files by using Adobe Acrobat Professional directly on a notebook, rather than using the "Save as..." feature in Mathematica. (At least on Windows.)
The CDF files don't allow the use of third party packages so I find them extremely limiting. Much research or advanced teaching would probably involve the deployment of private packages.
It is still my opinion that Mathematica itself ought to be the premier and widely used medium for technical communication. It has features, which if properly deployed, are orders of magnitude better than conventional publication mediums.
The reason that Mathematica documents can be so much better than conventional documents is because of their memory, ability for very general active calculations, the ability to provide highly customized dynamic presentations of data and concepts, and the ability to transmit active generated knowledge. If a practice of providing all needed calculation infrastructure and "calculating everything" is pursued then these documents will have much higher integrity than conventional documents because of the high degree of self-proofing that occurs. (One can still make errors but it takes quite a bit more work!) Using basic routines in graphics and dynamic presentations is a good way to expose problems. Such Mathematica documents are much more likely to foster collaborations because a reader will have the tools in his hand to vary, extend and verify.
Unfortunately (at least it seems so to me) WRI has only a hazy vision of this and are only fitfully stumbling toward its realization. They spend time on approaches that cannot fully realize the goal, or provide facilities for transfer to old mediums that are not adequate. One of the problems is their "top-down with options" approach rather than a bottom-up approach. This can't help but be biased toward old-style presentations and discourage experimentation and development of new and better presentations. Why in graphics isn't everything a graphic primitive? Why can't a user have more control in what goes around a graphic? I also believe it is a step in the wrong direction to have Mathematic automatically print suggestions and hints to the notebook. There are better ways to give access to relevant help - such as right-click context menus or specialized palettes with menus of routines and quick individual links to Help. A notebook should start as a blank sheet of paper on which the writer can write what he wants and not have somebody come along and add doodles on it.
Workbench documentation capabilities could also be significantly improved. It has rather languished for a number of years. Significant collaborative research projects could benefit from an Application format.
Finally, it takes a lot of experience in trying to write significant documents in Mathematica (a stand-alone Manipulate with a paragraph of text is not a significant document!) to determine the best forms of interface.
What saddens me is that WRI was getting so close and then seems to have veered away to a collection of special and poorly integrated applications. Someday everybody will develop, write, publish and collaborate in Mathematica. Someday.
From: Dr. Robert Kragler [mailto:email@example.com]
Hallo MathGroup readers,
all the nice features of Mathematica which allows to save even sophisticated nb files in formats such as CDF and pdf are (unfortunately) irrelevant when it comes to the point of preparing some documents for conference proceedings. For example, Springer Verlag always requires that articles are submitted in (La)TeX format which would require a lot of additional effort to retype formulas etc. and copy pictures separately into the TeX document. To my opinion, TeX was a great achievement in the 80-ies, but 30 years later it seems to me that this is some kind of stone-age IT.
In the case of producing a pdf file it requires (to my experience) only a printer driver e.g. such as "pdfFactory" which creates reliable pdf files from Mathematica notebooks directly. Doing the same with "Save as..." within Mathematica (V8) produce huge pdf files for which e.g. Adobe Acrobat 8 Professional crashes. Thus, I can highly recommend the software tool "pdfFactory" for this task.
Now, my question is : does anybody know a software tool, not necessarily a freeware program, (working in a similar way as a printer driver) to produce a TeX file directly from a Mathematica nb file (with refined graphics etc.) ? Any recommendations will be appreciated.