Was Markdown Possible in the 90's?

See topic title. Could Markdown have been done instead of HTML?

Takes a quick glance at markdown. lightweight markup language for creating formatted text using a plain-text editor. We have come full circle, from the text formating programs like troff under unix.
ASCII could have been defined with text formating control characters, rather than fighting over
international symbols like Pound vs Dollar, or stupid control characters like NACK. A data channel
with embedded control was outdated since the 50’s.torn tape
Ascii may already have some markup control characters, but nobody realy has explaned just what
all the control characters do.


And troff goes back to RUNOFF which takes us all the way back to 1964.

Hierarchical markup like HTML is much like Engelbart’s NLS, also from the 1960s.

1 Like

HTML was already kind of the Markdown of its day as a stripped down version of SGMLguid. Mind that containers weren’t a general idea in the beginning, so opening tags for things like list elements or paragraphs were totally sufficient and even the body tag was an optional luxury. With most tags consisting just of one or two characters, it was quick and easy to write.

I recall teaching a client, who had no previous exposure, basic HTML for formatting their texts in about half an hour. (This did not include tables or forms, which were then about the most complex things around.)

But there had been something, which was somewhat similar to Markdown: IPTC text, which was used for press releases. My first job was at the national press agency, and there was a Perl script, which took any files in a directory, scanned them, extracted headlines and dates, and formatted them as list with links to the respective documents. If called for a single document, it would translate the IPTC message to HTML, and it even had a search function for scanning the text portions.
I recall that the transformation included various levels of headings, paragraphs, lists, and even some sort of table. So while I don’t recall much of the IPTC specifications, this suggests that much what could be done with basic HTML could have been done with IPTC text, if you added links and embedded images. (Mind that IPTC text was specially crafted for ticker lines, so the encoding was pretty robust.)

Note: The format in question could have been IPTC 7901, but I’m not entirely sure, since this had been pretty much out of use for our purposes already, when I started to work there. (That said, IPTC 7901 is still used by news agencies today.)


I’m trying to figure out why things are the way they are. To Me Markdown feels simpler than HTML to write, but at the same time? That’s me thinking as an end user. Fewer keyboard motions to achive a thing ending i na document that is roughly human readable as raw text.

Apologies if this seems trivial. Heck, I learned HTML, or at least a subset, in the mid 90’s.

Much of this probably adheres to some sort of fashion. At that time SGML applications looked like a great idea for robust and universal formatting, so we got this entire HTML/XML/UML/SVG/etc path. Personally, I have a feeling that Markdown wouldn’t have been considered a serious enough format to serve the purpose. (Mind that these were the days, when every serious programming language had to have a mandatory statement delimiter, otherwise, it was for kids… :wink: )

As for basic HTML (up to HTML 3.2), I’d make a point in its favor for human readability and ease of use. (E.g., if you’re somewhat used to it, things like <i> are <b> are very fast to type and there is no ambiguity in what it may mean, as in “is bold one or two asterisks?” or “is this the start or the end of a formatting run?”.)

Edit: It may also depend on your perspective on what HTML actually is. If it is just a text formatting specification, Markwown is probably fine, as well. However, if it is a declarative programming language (which was often stressed at the time), you may have had a hard time convincing anyone that Markdown was a well-formed and sensible programming language. On the other hand, the container structure of SGML applications lent itself quite easily to the idea of object oriented programming, which was rapidly gaining ground at that time.


There were a couple of text markup languages from the late 80s that saw some use: texinfo (still used by GNU projects, strongly influenced by Scribe) and AmigaGuide. For a while, the gnu makeinfo command could create AmigaGuide documents, but it hasn’t been able to recently.

Markdown itself is absurdly similar to Setext (1991).

1 Like

I can’t find the article again but I remember reading something late 90’s that the creator of HTML had designed it with the notion that you would use a dedicated editor instead of writing it by hand.

I didn’t know this – thank you!

Personally, I think, the crucial element for the success of HTML was the introduction of distinctive and specific clear-text end tags. So you got unambiguous containers, which probably helped much regarding the acceptance as a somewhat serious format.
(I recall anything web-related being looked down at by “serious” IT folk, thus the claim “but it’s a declarative programming language”, and also “but it doesn’t have feet, it doesn’t run on the server”, at the same time, which may have been a strange concept to some. And of course, JavaScript wasn’t something a serious programmer could be bothered with. So anything that strengthened the chances of acceptance was a crucial factor in the tribal infighting over resources. Mind that only few players at the time started out as dedicated web shops or web distributors.)

1 Like

And yet another perspective: From a production/development point of view, HTML allowed you to come up with all sorts of ideas. You had a look a look at the specs and went, “what can I do with this, how can I trick this into behaving in that and that way?”, come up with a bright idea and check, if the major browsers played ball. So you got things like non-linear content presentation, tables misused for layout and design, creative uses of frames and backgrounds (yet another crucial element of HTML: extensibility by tag attributes), etc. I doubt that much of this — which contributed much to the attractiveness of the Web –– would have been possible with a format similar to Markdown. In the end, it wouldn’t have been much different from BBS presentations, which probably wouldn’t have been enough to attract broader audiences. But with HTML you got both aspects at once: simple formatting and presentation of linear text and fancy presentation. HTML somehow hit a sweet-spot of “unspefic specificity”.
So I think, you could have done similar with Markdown, but it wouldn’t have become the Web as we know it, and it would have probably stayed rather niche.

1 Like

Mind that Netscape browsers came with a WYSIWYG HTML editor (which also facilitated HTML mail formatting). If I recall this right, this was limited to the Gold edition in some iterations.

1 Like

I see texinfo as TeX mangled to look like Scribe. It should not be a surprise that Stallman wrote the EMACS Manual (AI memos 554 and 555) in Scribe.

And texinfo was of course also influence by INFO (the ITS version, GNU info is essentially identical). This goes full circle because INFO is much like the markdown languages we have right now.

1 Like

INSTEAD of, no. But that’s because it’s really kind of aimed at being an adjunct to HTML and similar rendering systems.

I see the usefulness of MD being that it does translate one way to HTML one to one, but does so in a format that is also purely text with no separate markup, so to speak. In HTML the markup makes the result difficult to read, the h1s and tts and such detract from the original text. In MD, the markup is what you would likely use in a pure text file anyway, so it remains readable in a native format as a pure text document. Most of the time anyway.

1 Like

Yes… I think I’d say (loosely) that markdowns are primarily human-facing, whereas HTML and related languages are primarily machine-facing.

I think it’s important that the system of tags and nested structures in the HTML family can be parsed out even if there are tags and structures from the future - HTML and the like are nicely extensible and backward compatible, in principle. (In practice, it became a big mess which will never be cleaned up, but that’s life.)

I’ve always been a fan of plain human-readable ASCII documents, so markdowns do appeal to me as a compatible extension of that. Ideally, a normal person with no knowledge of markdown should be able to read and comprehend the structure of a marked down text.

(Of course, ASCII isn’t enough, but then we get into encodings which is a whole other topic.)

1 Like