Open HughP opened 1 year ago
Greetings,
Reading the above suggestion made me think about it too.
JATS XML, if I understand it correctly, is just an XML schema for specific types of documents (and conversion to anything else relies on e.g. XSLT stylesheets). XLingPaper XML, I am not so sure, might be in the same case. So I am not sure these really belong here -- or any document-related XML schema would be suitable (say TEI XML, etc.)... It's kind of a different topic.
On the other hand, I do believe XSL-FO, with existing commercial processors but also an open source solution (Apache FOP) ought to be mentioned. FOP could be considered as a "a digital typesetting system (...) arranging text and media into digital documents". Yes, the "input" format is XSL-FO, which is not particularly human-friendly, but it is usually just a part of a broader solution. After all, FOP and Co are implementing line breaking, paragraph assembling, page layout, font selection etc., = the deep core of all typesetting systems. There are several markets using this approach to produce lengthy documents in PDF and HTML --- for instance, one could mention DITA, a specification focused on developing technical documentation (in XML syntax or in LwDITA which is a specialized lightweight Markdown syntax) is quite used in the industry for software and hardware documentation; it has dedicated tools, either commercial (there are several ones) or open source (at least two implementations to my knowledge, the "official" DITA-OT and XMLMind's DITAC). It's sort of a targeted niche market, but the specific challenges met by technical product documentation are usually forgotten in other typesetting solutions.... Both DITA-OT and DITAC rely on XSLT, for conversion to other formats (usually HTML), and on XSLT+XSL-FO processors such as FOP (for PDF generation).
Another solution which might be worth mentioning is Paged.js. Whatever one may think about ultimately delegating typesetting to a (headless) Web browser, it has some momentum and would tick several of the checkboxes here... The web stack is mentioned on the website (in "Separation of concerns"), so it's hard to avoid discussing it ^^
Regards. (Disclosure: I am currently quite involved in using SILE. I used DITAC a few years ago in my professional life)
Thank you for your response. Thanks for the pointer @HughP and for the details @Omikhleia!
If I write software which literally just replaces <body
with <body data-format="html5"
in some HTML5 file, I think it would also satisfy the criteria listed on the website (because they inherit the properties of HTML5). Or if we have some Javascript library, we have the same situation. But adding a <script/>
element really was not my intention of what a typesetting system shall be. My inclusion criteria are not mature yet.
I wonder whether a certain notion of conversion shall be included. Specifically about Paged.js, I don't think it should count as typesetting system, but it is not terrible if it fits the criteria. In general, a typesetting system should do some typesetting tasks like arranging elements (obvious when you have paged output documents) or conversion of structure/metadata/serialized data (typical minimal task of a system generating reflowable documents).
So maybe we could extend “3. Produces HTML5, PostScript, or PDF 1.4 (or later) output” to “3. Produces HTML5, PostScript, or PDF 1.4 (or later) output and its input format must not be the same”. Any opinions?
By the way, another addition could be Lout. It got somewhat hyped when it came into being, and was advertised a new (La)TeX too as some of the other solutions now are... it even has a Wikipedia page... I believe it's quite dead nowadays, or at best stuck in its time, but it had interesting concepts and it does tell something, also about the possible failure of such attempts. Will SILE, Typst, Patoline, etc. stand the test of time? Ah!
I feel like you should add JATS XML, XLingPaper XML, XSL-FO, TeXML, and Adobe Indesign.