Дата публикации: 03-10-2026 21:59:15

Yours is an interesting opinion. It comes to me a bit as playing the devil’s advocate, but playing the devil’s advocate is often a useful exercise. So, I am thankful for your comment.
I think you understood my point that having a disproportionately wide table makes the act of parsing the information contained in it very hard for any human. I think we are not disputing it.
Do you find that a table appearing like in this block
|Item |Description |Value|
|-----|----------------------------------------|-----|
|col 1|cell with very long text, which needs to|1000 |
┆ ┆be wrapped over multiple lines. Like in ┆ ┆
┆ ┆this example. ┆ ┆
|col 2|another line |2000 |
|col 3|and yet another |3000 |
would be hard for human to parse? I thought any human would immediately understand from the context, even without ┆ what cells are wrapped and what not – I expect that in 99% of the situations, understanding what is wrapped is pretty straightforward for any human.
My main concern and goal with the proposal above was to provide a format understandable by machines, so that they know how to export the table into different formats and how to reflow cells split over multiple lines into a single-line cell. I wanted this process to work without any parsing of the content. It should be just autoamtic for a machine based on simple rules.
I really appreciate your thoughts and comments, but as of now, I am not persuaded by your criticism.
That being said, I am not attached to my proposal in any way. I only wanted to revive the discussion around multiline cells. I think it would be wonderful if there were a wide convergence of many people towards a format. I don’t mind if it is a standard or not. Apparently, Markdown has worked mostly the other way around. Things have been adopted by many people and only later standardized.
So, if you say the pandoc multiline_tables is better, I have no objection to adopting it. The problem is that I saw no one so far using it (but maybe it is just my limited perspective).
So, does pandoc propose to leave a blank line between table rows with some delimiters for starting and ending the tables? The only problem I see with the format is that the GFM dialect is way more widespread. Is it possible to take from pandoc’s multiline_tables the idea of blank lines and import it to the GFM dialect? I am not so sure because a blank line always meant that the table ended there.
So, I am generally open to discussing any other idea, as long as it is a simple format and can reach a wider consensus.
A small aside, which is likely not needed as I presume we share the same opinion: What I am not interested in are formats that aim to include too many features like cells spanning over multiple rows and columns. These are interesting proposals, but they are, in my view, too narrow. If the same information repeats over multiple cells, one can simply write idem to imply that the information is repeated. For me, markdown needs to remain a pure text format, simple to write and simple to read. Its primary mission is not perfect, typography-quality typesetting. If this high typographical quality can be achieved by exporting to a different format like a LaTeX table, even better, but it should not be the guiding principle.

Основное содержимое страницы с новостью.

Схожие новости