jdev - 2021-11-13

xml namespaces is one reason I don't like using xml much, conceptually having just a second nested element if neccesary would work too, but namespaces have yet another way to represent the same conceptual data in it : /

Namespaces aren't quite the same as just having another element, because they allow you to identify elements and prevent conflicts when people extend

They are not exactly the same, no, but conceptually there is not that much difference

It's nice knowing where some features in a stanza come from

Do you know how hard it is to get search engines to find what you're looking for?

Especially if you don't know the name of something

i hate when the data you search is the name of the tag

like in https://xmpp.org/extensions/xep-0108.html

i find that bad design

under the assumption that every element is exendable, this means i have to iter the childs, then compare the found tag name against some kind of allowed list

treat it as an enum?

yes but thats actually not what i want in an extendable protocol

someone maybe wants to introduce a new element

long story short, this makes things unnecessary error prone

instead do <activity> <general>relaxing</general> <specific>reading</specific> </activiy>

this way i dont need to iterate through childs, i can just tell, get me that element text

it does not matter what other client stuff into this, it will never alter my way of finding what i need

and its perfectly extendable if someone wants to put some new general activity which is not in the list the XEP defines, i can still take it and present it as an activity

I know what you're saying, but that's also harder, you can't just stick arbitrary text in there to "extend" it

on the other side if i find a new child which is not in my list, i dont even know if the goal was to extend the lists of activitys, or if this is some completely different shit

You wouldn't get an unknown child in the same namespace ✏

That's the point

you mean everything in that namespace is an activity

Everything defined as an activity, yes

.....

Extending it would use a new namespace (as shown in examples)

thats not the same thing

<activity> <relaxing/> <notanactivity/> </activity>

That's illegal

That would be an error.

i can treat this as an protocol violation?

Yes

ok thats good

that makes things more sane

Just get the first child of <activity> that's in the XEP's namespace, it will always be one of the ones defined in the XEP

If not, it's invalid

Ignore stuff in other namespaces

sadly not MattJ, because there is also a <text> chid

the xep forces you to Iterate

Something like `getChildSetNS(Set<Name>, Namespace)` would be nice

Unfortunate

from a implementor view, i want things i want to query a specific thing, not go start searching for something

lovetox> i find that bad design it is bad design

What's a better design?

i put both into attributes

then you can still extend the <activity> tag with childs in other namespaces if you want

or as i said before, define elements and set the values as data of them

i dont understand why the custom extension which is an example in that xep needs to be a child of a specific activity

like its a custom extension anyway not defined anywhere so what does it matter if its a child of a specific activity or just the <activity> tag

The idea is that it's not an additional activity, but a more specific one of the same type

What if you're doing 3 things at once!

So if you understand the specific one, you can use that. If not, you still see the outer one

MattJ, simply move the name of the activity into an attribute would work then

<activity xmlns='http://jabber.org/protocol/activity'> <general name="inactive"> <specific name="sleeping"> <hibernating xmlns='http://www.ursus.info/states'/> </specific> </general> </activity>

i would prefer that probably

<activity xmlns='http://jabber.org/protocol/activity' general="inactive" specific="sleeping"> <hibernating xmlns='http://www.ursus.info/states'/> </activity>