Markdown is a beautiful demonstration that document structure syntax can/should be simple. What most people do in Word is better done by just adjusting the document rendering/style, not the document structure...
I love the idea of extending markdown to include more visual elements, but if you're not careful you just reinvent HTML.
Here's my personal take on extending table syntax for charts. Easy to write, and if a renderer/parser understands the syntax you get a beautiful chart, and if it doesn't you get a table with slightly weird headings:
Tables are the one thing in markdown where I’d prefer to emphasize edit ergonomics over good looking unrendered text. Making a quick manual change like adding column to a markdown table is just unfun. I’ve always thought a json like format that a linter can organize would be better.
Which is all to say I really like the table proposal here - adding an optional linter to make the data look tabular in unrendered markdown will make it even better
Making a quick manual change like adding column to a markdown table is just unfun.
This is one of those moments where I realize that the vim life spoils me. It's so easy to do this in vim that I don't even think about. I probably use it a dozen times per day such as commenting out code.
Ctrl + v, select where you want the character, then hit I (shift + i), type your thing, hit escape, and Bob's your uncle.
Even in Vim, the editing experience falls over when making markdown tables that have non-trivial content in their cells (multiple paragraphs, a code block, etc.). I recently learned that reStructuredText supports something called "list tables":
Where a table is specified as a depth-2 list and then post processed into a table. Lists support the full range of block elements already: you can have multiple paragraphs, code blocks, more lists, etc. inside a list item.
This syntax inspired the author of Markdoc[1] (who came from an rST background) to support tables using -separated lists[2] instead of nested lists (to provide more visual separation between rows).
I have found various implementations of list table filters for Pandoc markdown[3][4], but have never gotten around to using any of them (and I've tossed around ideas of implementing my own).
reStructuredText & AsciiDoc are so, so much better than Markdown since they have rich feature sets to actually build documentation, blogging, & so on. It’s a massive shame everyone would prefer _yet another Markdown fork_ like the OP.
Not to defend Word et. al. too much, they have plenty of problems, but keeping the structure simple and applying a style over it is a completely supported way of doing things.
I have documents with essentialy zero direct styling, just paragraph styles (for headings, bullets, code blocks, quotes) and character styles (links, inline code). The UI isn't super well optimized for that, but once you get used to it, it's so much nicer than Markdown or LaTeX for multi-page print work.
I work on a dashboarding / BI solution that is also built around markdown and clickhouse. www.evidence.dev
We moved to stripe's Markdoc variant for the component syntax last year and have been really happy with it. Models are good at writing it, people are good at reviewing it.
Here's an area chart that would issue a SQL query for weekly revenue totals:
My guess is parsability. It’s easier to look for sentinel ``` blocks as opposed to building an HTML processor. An XML processor would have been easier, but people like Markdown. So, here we are.
I also went with Front Matter for styling and added an interactive styling mode you can do on the web to test it out immediately. There are some examples on my homepage which demonstrate it in action.
SDocs is cli -> instantly rendered on web
Despite being in the browser, the content of SDocs rendered Markdown files remain local to you. SDoc urls contain your markdown document's content in compressed base64 in the url fragment (the bit after the #):
https://sdocs.dev/#md=GzcFAMT...(this is the contents of your document)...
The url fragment is never sent to the server (see https://developer.mozilla.org/en-US/docs/Web/URI/Reference/F...: "The fragment is not sent to the server when the URI is requested; it is processed by the client").
The sdocs.dev webapp is purely a client side decoding and rendering engine for the content stored in the url fragment.
This also means you can share your .md files privately by sharing the url.
Also, I’m sorry I high jacked your post to some degree with this comment.
It’s just a little too relevant for me not to leave a comment!
I'm starting to wonder if someone is selling 'open source CV booster' packages to executives looking to jump ship. This is obviously vibe-coded with Claude, has a broken link to /superpowers which indicates no human has read the README before publishing.
The profile seems to be a real person, CEO/CTO level of a small/medium company, but with zero history of OSS contributions, publishing anything or social media presence. It's the third project of this kind to show up in HN this week.
Broken /superpowers link was a real miss — I removed that directory because it had private context and didn't audit the README for references. Fixed now. CLAUDE.md is public in the repo if you want to see how it was built. Not trying to sell anyone anything; it's an experiment I put out to see if it's useful.
Indeed. This is just a vibe-coded addition to an already overcrowded space, with no indication of any intent to consult others or support it for real use.
I was expecting to find a link to a github pages site where I can see the rendered examples, but only found a link to the html sources in examples/out. Am I missing something?
jesus thank you, im so wary of any project that isn't going to do this bare minimum. static pages on gh are literally free it feels absurd to post this project to HN without doing that.
I feel your pain. My project is https://sdocs.dev, the homepage is actually the rendering of a markdown file (sdoc.md), so you can see how SDocs renders immediately. There are some links on the homepage to other Markdown files with some elaborate styles and charts
nope this is becoming a theme. many showhn posts that are about visualizations do not have any renderings of said visualizations. especially in github readme.
- add lint or errors, otherwise your formatting will break, e.g. LLMs and humans will add text too long or too short and your design system will not be able to handle this.
- it's great for low token input
- validate the layout of the user vs. the components used.
You have gone the full latex route. very interesting project. my purpose was simple, to keep mdv extremely simple nothing complex. I do not want full html/latex replication and for surely no inline code...
The only reason to use markdown is that you can read it and write it in a text editor unaware of syntax and rendering semantics. When this purpose is lost and your UX or reading or writing a document starts depending on renderer and your knowledge of how it works, the result becomes just another rich document format, where editor abstraction is a must. And then why all this is even needed? ODF does the job well.
Nice project. But at what point does Markdown just become Emacs Org-Mode? At least with Emacs you can write Lisp to make your document do anything you want.
55 comments
I love the idea of extending markdown to include more visual elements, but if you're not careful you just reinvent HTML.
Here's my personal take on extending table syntax for charts. Easy to write, and if a renderer/parser understands the syntax you get a beautiful chart, and if it doesn't you get a table with slightly weird headings:
Which is all to say I really like the table proposal here - adding an optional linter to make the data look tabular in unrendered markdown will make it even better
>
Making a quick manual change like adding column to a markdown table is just unfun.This is one of those moments where I realize that the vim life spoils me. It's so easy to do this in vim that I don't even think about. I probably use it a dozen times per day such as commenting out code.
Ctrl + v, select where you want the character, then hit I (shift + i), type your thing, hit escape, and Bob's your uncle.
https://docutils.sourceforge.io/docs/ref/rst/directives.html...
Where a table is specified as a depth-2 list and then post processed into a table. Lists support the full range of block elements already: you can have multiple paragraphs, code blocks, more lists, etc. inside a list item.
This syntax inspired the author of Markdoc[1] (who came from an rST background) to support tables using
I have found various implementations of list table filters for Pandoc markdown[3][4], but have never gotten around to using any of them (and I've tossed around ideas of implementing my own).
[1] https://markdoc.dev
[2] https://markdoc.dev/docs/tags#table
[3] https://github.com/pandoc-ext/list-table
[4] https://github.com/bpj/pandoc-list-table
I have documents with essentialy zero direct styling, just paragraph styles (for headings, bullets, code blocks, quotes) and character styles (links, inline code). The UI isn't super well optimized for that, but once you get used to it, it's so much nicer than Markdown or LaTeX for multi-page print work.
I continue to love Markdown and always push it a bit further than Commonmark, with frontmatter, schemas, code fence metadata too.
I've been enjoying https://djot.net/ as a superset of Markdown that is feels very well designed and extensible too.
You may look into its syntax and tooling for prior art or some extra lift.
I'm trying to get a djot extension in Zed for syntax highlighting if anyone minds adding a to help signal some community interest.
https://github.com/zed-industries/extensions/pull/5206
We moved to stripe's Markdoc variant for the component syntax last year and have been really happy with it. Models are good at writing it, people are good at reviewing it.
Here's an area chart that would issue a SQL query for weekly revenue totals:
``
{% area_chart data="my_table" x="date" y="sum(revenue)" date_grain="week" /%}``At least, that would’ve my rationale.
This problem has risen to the top of many people’s minds at this moment (including mine!). My Show HN for a similar cli + web based solution (https://sdocs.dev) is on the /show page now (https://news.ycombinator.com/item?id=47777633).
I also went with Front Matter for styling and added an interactive styling mode you can do on the web to test it out immediately. There are some examples on my homepage which demonstrate it in action.
SDocs is cli -> instantly rendered on web
Despite being in the browser, the content of SDocs rendered Markdown files remain local to you. SDoc urls contain your markdown document's content in compressed base64 in the url fragment (the bit after the
#): https://sdocs.dev/#md=GzcFAMT...(this is the contents of your document)... The url fragment is never sent to the server (see https://developer.mozilla.org/en-US/docs/Web/URI/Reference/F...: "The fragment is not sent to the server when the URI is requested; it is processed by the client").The sdocs.dev webapp is purely a client side decoding and rendering engine for the content stored in the url fragment.
This also means you can share your .md files privately by sharing the url.
Also, I’m sorry I high jacked your post to some degree with this comment. It’s just a little too relevant for me not to leave a comment!
> Can you tell me more about the :: blocks thing. I didn’t know that was a Markdown element.
They are Markdown directives
/superpowerswhich indicates no human has read the README before publishing.The profile seems to be a real person, CEO/CTO level of a small/medium company, but with zero history of OSS contributions, publishing anything or social media presence. It's the third project of this kind to show up in HN this week.
> .mdv is strict CommonMark plus four additions:
> YAML front-matter for title, theme, named styles, and dataset references.
> Fenced blocks for data/visuals: ```chart type=bar x=region y=sales.
> ::: containers for styled regions and layout: ::: callout / ::: columns.
> ::: toc for an auto-generated table of contents.
https://wire.wise-relations.com/guides/components/
my takeaway:
- add lint or errors, otherwise your formatting will break, e.g. LLMs and humans will add text too long or too short and your design system will not be able to handle this.
- it's great for low token input
- validate the layout of the user vs. the components used.
- seen here before: https://myst-parser.readthedocs.io/en/latest/syntax/optional...