💾 Archived View for circumlunar.thebackupbox.net › ~epoch › blog › text_form captured on 2024-12-17 at 10:28:33. Gemini links have been rewritten to link to archived content
⬅️ Previous capture (2023-01-29)
-=-=-=-=-=-=-
I didn't know w3 XForms existed, but ofc it is 10 seconds long measured by holding page-down on my keyboard.
but I also found out about:
how I found out about these two things. thanks, theruran.
I guess content-type would be:
text/form
need a way to have a list of inputs.
each input should have a type associated with it
a name that is mainly for machine use
a description that is only for human use
maybe a label? description and label might be similar enough to merge.
input:
type:
text
name:
user
label:
username
desc:
your username on thebackupbox.net
input:
type:
password
name:
pass
label:
password
desc:
the password that goes with your user on thebackupbox.net
uh..
I dunno. there's probably something that has a good list that has thought about this more than I have.
looking into semantic web stuff atm...
each form needs an action URI...
form:
action:
http://thebackupbox.net/cgi-bin/fmrl-client.cgi
input:
type:
text
name:
user
label:
fmrl user
description:
the fmrl user of which you wish to view updates
default action is to just launch the action URL with the inputs appended in the usual way in a query_string
name=value pairs separated by &s or ;s for multiple inputs.
this /might/ need a method attribute for the form.
maybe http:POST would be good for what the value of this attribute would look like?
the two main types would be:
maybe there should be methods that manipulate the URL in other ways.
like something for appending the input to the path?
like, ftp might do put.. http doing POST or PUT..
sftp doing put as well.
as with HTML forms, file inputs probably shouldn't get included as parts of the query-string-generator method.
I guess it /could/ be done as json if each input was part of an array so the order was retained.
like:
{ "form": {
"action" : "https://thebackupbox.net/cgi-bin/login.cgi",
"method" : "http:POST",
"input" [
{
"type": "text",
"name": "user",
"label": "user",
"description": "long string"
},
{ "type": "password",
"name": "pass",
"label": "I dunno.",
"description": "just an example"
}
]
}
}
or...
<xml>
<form action="https://thebackupbox.net/cgi-bin/login.cgi" method="http:POST">
<input type="text" name="user" label="user" description="long string" />
<input type="password" name="pass" label="I dunno"." description="just an example" />
</form>
</xml>
I guess if I'm going to do XML-based I'll write a DTD for it...
could also be in yaml? which I guess would be the examples I did up there.
or maybe toml
https://en.wikipedia.org/wiki/Recfiles
rec file would work too.
...
I've worked my way to RDF and Turtle and N-Triples.
help.
I'm trapped in semantics.
so far this is just taking pretty much HTML forms and stripping the HTML off.
I figure the actions should be influenced on what curl does.
it does a lot
comments: