Re: [xml-dev] The impact of data format selection on applicationdevelopm

["Liam R. E. Quin" <liam@fromoldbooks.org>]

On Tue, 2022-07-12 at 11:51 +0000, Roger L Costello wrote:
> Hi Folks,
> 
> I have been having a discussion with a colleague on this topic. I
> thought it would be useful to share part of our discussion. 
> 
> [To remind people what the topic is about, we are trying to come up
> with rules/heuristics on when to use super-simple data formats versus
> when to use comparatively complicated data formats such as XML or
> JSON.]

"Simple" is both contextual and subjective.

To many people, a Microsoft Word file is a really really simple format
for exchange.  Microsoft Word never presents them with syntax errors in
the format while they use it!

But to a programmer, it's a very complex format.

Similarly, markdown and wiki syntax are very simple for simple things,
to people who are OK with editing text files. But if you've ever edited
a table in a wiki or with markdown, you'll know that the format stops
being simple when you get past its primary focus.

If you've ever wanted to nest one table inside a table entry in another
table in a wiki or with markdown, you've probably decided pretty
quickly that the root canal surgery down the corridor is a better
option.

For a programmer, markdown is a often case of using a library to
convert it to HTML or JSON.

There's no simpler format for a JavaScript programmer than JSON, but if
you try to edit a half-megabyte JSON file in a text editor you're back
at the dentist - that's not what JSON is for.

The Unix file format of lines terminated by a newline was a breath of
fresh air for system integration and scripting at the time, especially
when combined with the regular syntax for filenames (or, originally,
nodes in the graph). But is the order in which the lines appear
significant? And many programs extended the format - troff, for
example, had a very complex syntax, used for typesetting, and any
language in which you might need to type sixteen \ signs in a row can't
be called simple :) - but troff fit in to Unix because of the pre- and
post- processors that did tables, equations, diagrams, and that
supported different output devices.

A format is simple when it meets the needs of its users without
unnecessary complication or emotional trauma.

There is rarely any benefit in using a format that is not sufficiently
powerful to meet your needs.


liam

-- 
Liam Quin, https://www.delightfulcomputing.com/
cancer fund - https://www.gofundme.com/f/5u9v7-every-little-helps