<h1>Why Submit the TeX/LaTeX Source?</h1> <ol> <li> TeX has many advantages that make it ideal as a format for the archives: It is plain text, it is compact, it is freely available for all platforms, it produces extremely high-quality output, and it retains <a href="#contextual">contextual information</a>.<br /> It is thus more likely to be a good source from which to generate newer formats, e.g., MathML [namely HTML, or more specifically XML, that handles mathematics correctly -- note that the MathML people plan a LaTeX to MathML translator, but dvi/ps/pdf lack the necessary document structuring concepts]. Possession of the source thus provides many additional options for future document migrations. </li> <li> Using emerging new technology, most of the Postscript generated from the source contains <a href="/hypertex/">hyperlinks</A>, so that using new versions of PostScript previewers (e.g., the next version of Ghostscript/Ghostview), readers can point-and-click to navigate within the paper, and even over the web itself.<br /> And by archiving the source, we maximize the potential for seamless adoption of future technological improvements. Archived papers can be repeatedly rejuvenated by automated reprocessing. </li> <li> We distill the source into Adobe's PDF format. The hyperlinks can then be <a href="/help/pdf">viewed using Acroread or an equivalent reader</a>. In addition, this means that the hyperlink overlay will be directly available to those web browsers with built-in PDF support (thus combining the network transport capabilities of html with a more serious page markup language). Having the paper available in PDF means that problems discussed below with Mac and PC browsers can be avoided. </li> <li> There is no single Postscript standard! We provide Postscript in many formats; this is not possible if the author submits a single Postscript file. For a historical example, Preview.app under NeXTstep displayed bitmapped fonts poorly, and for historical reasons (TeX predated common use of Postscript) these fonts are still most commonly used by TeX and dvips. By requesting type 1 fonts from the "More options" page, screen readability is improved. </li> <li> Cross-referencing within arXiv is added automatically with hyperlinked Postscript. Authors should specify <a href="/help/arxiv_identifier"><b>arXiv:YYMM.NNNNN</b></a> or the older format of <b>arch-ive/papernum</b> (e.g., hep-ph/9503456) references whenever available, and these strings will be pattern-matched and replaced with suitable hyperlinks back to arXiv. Similarly, any occurrences of <b>http://...</b> or <b>ftp://...</b> url's are detected and converted to hyperlinks. </li> </ol> <hr noshade /> <a name="contextual"></a> <h2>What is "Contextual" Information, and Why is it Important?</h2> <p> We mean by this the relationship between equations and their labels, references and their numbers, subsection headings and their entries in the table of contents, and so on. While ordinarily readily available in TeX/LaTex source, conversion to Postscript irretrievably loses this structural information. The loss is unfortunate because with new formats such as PDF, the information can be used to provide active hyperlinks: e.g., in a PDF viewer you can click on an equation number and jump back to the specified equation. Moreover TeX itself can be processed as HyperTeX and, with the proper dvi previewer, clicking on equation numbers will bring up the desired equation in a separate window, or even retrieve other papers specified by their proper arch-ive/papernum identifier. HyperTeX works by redefining the standard macros and works retroactively for pre-existing TeX/LaTeX source -- HyperTeX conversion is accomplished by merely re-TeXing with the modified macros. Since information is ordinarily lost in each stage of processing, TeX source contains (close to) the maximal amount of contextual information that can be retroactively processed into any future format which can take advantage of it. For more information, see <a href="/hypertex/">https://arxiv.org/hypertex/</A>. </p> <hr noshade /> <h2>Why doesn't arXiv Accept Preprocessed Postscript Submissions?</h2> <ol> <li> Postscript from TeX usually contains bitmapped fonts at a fixed resolution. Your favorite resolution is inappropriate for other users and may make your paper difficult or impossible to read. We will make available many versions of the Postscript including 300 dpi, 400 dpi, or 600 dpi bitmapped fonts, as well as Postscript with (un-embedded) Type 1 fonts.</li> <li>Different dvi to PS drivers produce Postscript that ranges from completely non-compliant with Adobe's Document Structuring Comments to only marginally compliant. We will always have the latest dvi to PS software installed and will produce DSC compliant PS.</li> <li>Postscript loses contextual information that is implicit in the TeX. This information is required to provide hyperlinked overlays to papers (so that users with the proper viewers can, for example, click on a reference number and jump to the reference [which can also include a web URL which can be automatically passed off to a web browser]). To do the hyperlinking, we use HyperTeX macros. The Postscript we produce is HyperPostscript which can be distilled into Adobe's Portable Document Format (PDF). Our software makes web URL's out of all references of the form YYMM.NNNNN or arch-ive/yymmnnn (e.g., 1510.00322 will be converted to a hyperlink for https://arxiv.org/abs/1510.00322, hep-th/9511053 will be converted to a hyperlink for https://arxiv.org/abs/hep-th/9511053).</li> <li> TeX source is more compact and space-efficient.</li> </ol> <hr noshade /> <h2>Frequent Red Herring Concerns:</h2> <ul> <li><b>Will the auto-TeXing embed figures?</b> <p> Yes. Our TeX installation can do anything that yours can. Any macros you use that we do not support can be included in your submission (or sent to us separately so we can put them on-line). </p> </li> <li><b>Won't TeX source make it easy to plagiarize?</b> <p> There is no file format or other technological device that can protect you from this. At the very least, unscrupulous re-typers would always remain a threat. Postscript does not provide a barrier in any event: it is quite simple for someone with a little knowledge to extract any text from a Postscript file. Moreover a plagiarist who cuts-and-pastes directly from your TeX source is all the more easily detected, since the source is easily identified. We archive all versions of papers so that we can assist in any priority or plagiarism disputes. </p> </li> <li> <b>I worked hard to make my figures and I don't want people to steal them. Shouldn't I hide them by embedding them?</b> <p> As with the above question, it is quite easy for someone with a little knowledge to extract anything they like from the output PDF or Posctript file. Furthermore, unauthorized or un-attributed use of figures counts as plagiarism, just as above, so the rest of the above discussion applies here as well. </p> <li> <b>What if my TeX source has potentially embarrassing self-comments in it?</b> <p> Well... you should probably take them out. It is easy to strip these out in advance of submitting. Here is a Perl filter to do it. Please, please do not hurt yourself with this script; save your file and do not write over the backup copy... just in case. <blockquote><code> #!/usr/local/bin/perl<br /> while(&lt;STDIN&gt;){ s/^\%.*$/\%/; s/([^\\])\%.*$/\1\%/g; print; }<br /> exit(0);<br /> </code></blockquote> <p> or use the one line command </p> <blockquote><code> perl -pe 's/(^|[^\\])%.*/\1%/' < old.tex > new.tex </code></blockquote> </li> <li> <b>I use the Textures program. Won't the archive destroy my paper's beautiful formatting?</b> <p> We have many Textures submissions here. Figures will, of course, be placed exactly where you put them (why would you expect otherwise?). Textures does use a non-standard command to control the way figures are included, please read our <a href="textures">notes on submitting Textures generated papers</a>. </p> </li> </ul>