Frequently Asked Questions

1. Can I submit Postscript instead of LaTeX?

   Sure, but LaTeX or TeX is much preferred for the following reasons:

   1. LaTeX is more portable than Postscript, and is in any case readable by those without access to anything beyond an ASCII terminal.

   2. The Postscript generated from the LaTeX is put up on the server, so that both LaTeX and Postcript are available if LaTeX is submitted, but only the Postscript is available if that's what is submitted.

   3. LaTeX is typically more space-efficient than Postscript.

   4. LaTeX can be searched more effectively, so that in the future, when we get full-text search capability, the LaTeX papers will be retrievable but Postscript papers won't.

   5. But most importantly, extended TeX and LaTex macros have recently been created that will automatically (and retroactively) turn TeX papers into HyperTeX that will allow the papers to interact with the WorldWideWeb. For instance you will be able to click on a reference number in a paper and jump to the references page and then perhaps retrieve that paper, and then jump back to where you were reading. Details to appear soon!

   For all these reasons, LaTeX submissions are in the long run more accessible than Postscript ones. If you absolutely must submit Postscript, please compress it and uuencode it first.

   Here are some of the standard excuses for not submitting LaTeX:

   • My LaTeX file includes a lot of comments to myself that I would rather not distribute.

     The following UNIX command removes comments from a TeX file:

     sed -e '/^%/d' -e 's/[^\\]%.*//' file.tex > newfile.tex
     

     You may want to check the output with diff to make sure the changes are appropriate.
   • It's a pain to put together all the files to mail them.

     The uufiles script, available from the server (`get uufiles') does everything for you automatically. It prompts you for a directory and set of file names, and generates a uuencoded, compressed, tar file of all the files ready to be mailed off to bayes-an.

     And anyway, you can't spare a few minutes to prepare something that over 750 people will have immediate access to?

   • People will just pull over the Postscript anyway, once that's on the server.

     Maybe so, but if the Postscript doesn't work, they'll still have a shot at using the LaTeX, which is more portable. Furthermore, we will soon be able to generate hypertext Postscript for submitted papers, which will only be available for those papers for which we have LaTeX.

   • I use a bunch of style files and personal macros.

     The style files can be archived at the server. Just package them up with uufiles and mail them in the body of a `comment' command. As for your personal macro file, why not just throw it in with the paper when you use uufiles to package that up?

   • Well, in any case, it's too late now. I already submitted Postscript.

     You can always replace the paper with the `replace' command.

2. I pulled over a paper that uses a macro package that I don't have. What should I do?

   The server archives standard macro packages and style files. A description of the archived packages is available here or by sending a `get macros' command to the server. Individual macro packages are accessible here or by sending a `get <macro-package>' command to the server, where <macro-package> is the name of the macro package you want.

   If the server doesn't archive the macros, send an email message to the submitter of the paper asking them to submit the macro package or revise their paper to be self-contained.

3. I want to submit a paper, but it uses some style files that aren't archived on the system. Can you add them to the macro archive?

   Yes, we can. You can submit standard macro packages and style files for archiving by sending them to us with a `comment' command. If you use your own idiosyncratic macro files, though, please move the macros directly into the LaTeX source file to make it self-contained.

4. You know, you should really have a WWW/Mosaic interface.

   We do. You're looking at it.

5. I downloaded a paper using the WWW/Mosaic interface, and I stored it in a file, but there's a bunch of junk at the beginning and end of the file. Is Mosaic broken, or did I do something wrong?

   Neither. As explained below (see the next question), multipart papers are sent out after being tarred and compressed, that is, as .tar.Z files. The typical vanilla client will uncompress them automatically, but won't untar them, hence the binary header junk.

   For more local configuration information and a sample script (with instructions) to automagically untar and run TeX, click here.

6. I requested a paper using the WWW/Mosaic interface, and I stored it in a file. What kind of file is it?

   Downloading the source (src), in the case of a single file, delivers as a compressed file (.Z), with mime

   Content-type: application/x-eprint
   Content-encoding: x-compress
   

   and, in the case of multiple files, delivers as a tarred compressed file (.tar.Z), with mime

   Content-type: application/x-eprint-tar
   Content-encoding: x-compress
   

   Downloading the Postscript (ps) delivers as a compressed file (.ps.Z), with mime

   Content-type: application/postscript
   Content-encoding: x-compress
   

   WWW clients that are correctly configured will automatically uncompress upon receipt, and pipe Postscript (when available) to a postscript previewer, or allow in the case of x-eprint or x-eprint-tar a user-defined pipe (via .mailcap) to a suitable processing script.

   For one such csh script together with instructions, try this.

7. I get way too much mail, so I'd like to get listings less often. Can this be arranged?

   The server sends at most one message a day, and it can be quickly perused. It's really an efficient information channel. However, if you really want less frequent messages, you can set this up yourself. Instructions are available on setting up a single monthly mailing with all of the title/author information for the past month by clicking here or sending a `get monthly' command.