Find a stupider data exchange format. Oh, that's right, you fucking can't because XML is peak stupid

Find a stupider data exchange format. Oh, that's right, you fucking can't because XML is peak stupid.

Attached: xml.png (523x480, 23K)

git gud

Depends on the kind of data.
Web API? JSON
Structured document markup? XML

.doc

>stupider data exchange format
json

it's usable but iso added too many stupid features like external entities

>added
More accurately, didn't remove from SGML.

We have a winner.

ASN.1
HTML or other frankensteined variants of XML (which tend to make it not context-free anymore)
BSON

the fuck is wrong with ASN.1

> XML bad
> LISP good

You will unironically defend this:

data

At least I can have comments, sane multiline support, and tools for validating and enforcing schema.

>at least i can have bloat, bloat, and some more bloat

Attached: acomputer.jpg (320x320, 24K)

Parsing it is a nightmare.
Especially when a schema references an external schema.

Omg I have never heard about BSON.
Now I looked it up....I wish I did not.

why wouldn't you like to have any of these in your data format?

The point of XML is self-describing data. That's not "bloat" thats the primary feature. The fact an industry of brainlets misuse XML where it's not appropriate doesn't excuse your misunderstanding.

XML verbosity is crap, but my pet peeve is "attribute or child element?" I don't think attributes should exist.

I bet you even comment your code.

all recursively enumerable ones
most context-sensitive ones
anything with ambiguous grammar
anything with multiple competing standards
anything with bazillion standard extensions
anything where implementation is the specification

XML isn't _that_ terrible

>attribute or child element?
The supposed rule-of-thumb is "elements for data, attributes for metadata". What counts as "metadata" is specific to your schema, but things like "id" for cross-referencing an element are usually attributes.

Like in this example (S1000D XML):

Replace the nozzle
Perform the following steps:


"id" (cross-referencing anchor), and "applicRefId" (reference to the context in which the step applies for filtering) are considered metadata about the step, while "title" and "para" are considered its actual content.

I never understood why it is
text
instead of
tag(var = x[){text}/code]

ah shit well i sure fucked up that post
tag(var = x){text}

Agreed. We should all upgrade our systems to exchange data encapsulated in .xls spreadsheets.

This

.docx

How is JSON, stupid? It's an easy to read format that contains nothing but arrays and dictionaries.

shut up fag

.ini
you lose

use protobuf niggers

Literally a glorified .txt

The whole web should be based on xml, but you fucking retards blew it.

Everything besides s-expressions from LISP is stupid

.pdf

Easy
INI Files

This is the Jow Forums equivalent of a reddit comment.

Yaml

So you don't end up with
}
}
}
}
}
but something more obvious like




When you use XML mainly for structured document markup and work with people of varying CS skills, the explicit end tags can help a lot.

>Structured document markup? XML
This. It's actually quite good at this, especially things which have quite a bit of human-readable text. I have some code-generation shit with is defined with XML and includes human-readable documentation right in the spec (which gets added to C header files), and it honestly works pretty well.
It fucking sucks as a configuration or general data-serialisation format.

Just because you don't get why it's a thing doesn't make it stupid you fucking brainlet.

Attached: 9149940-6645881-image-m-46_1548796680269.jpg (634x515, 37K)

Yaml

nah it's pretty good for conf files that have data in them

Too disgustingly overcomplicated.
Not the XML is any better.