Website without protocol

[Schlagi123]

Hi,

currently in JOSM URLs without protocol are valid (e.g. website=www.duckduckgo.de).
In the ​OSM-Wiki for the key website, this is not allowed. The protocol is not a optional part.
I think, we should generate a warning for such a URL and can also provide a fix.

[Schlagi123]

Patch for the issue

[taylor.smock]

It looks like this was allowed due to #10696.

iD issues:

• ​iD/issues/6831
• ​iD/issues/5866

iD code (commented out): ​modules/validations/invalid_format.js

Besides that, comments on your patch:

• var is a Java 11 feature. At this time, we are still targeting Java 8 (see #17858 for progress on switching to Java 11+ -- we may end up going from Java 8 to Java 17).

I'd have to look into how the rest of the code is affected by your patch, but if I were to merge the patch, I'd like to see a test for the fix command. But realistically, we could not automatically fix the problems, as we'd want to properly set http or https, which means we would run into the same issues as iD with respect to the autofix (do we really want to ping random internet servers? Probably not.)

Having an autofix for mailto seems OK.

EDIT: The wiki page was ​changed in 2015. The change has been there for awhile. So the rationale behind #10696 may no longer be relevant.

[stoecker]

Hmm. I'm not really convinced. Somebody simply removed the fact that this entry can only consist of a domain in the history of the wiki page. From a users point of view website often are stated without protocol in the real world. It's easy to handle this programmatically (each webbrowser does this). So I see no big sense in enforcing this (especially having the recent change from ​http:// to ​https:// in mind - maybe soon we have one to wwwng://?).

​Talk:Key:website#We_shouldn't_encourage_crippled_URLs
​https://wiki.openstreetmap.org/w/index.php?title=Key%3Awebsite&action=historysubmit&type=revision&diff=840167&oldid=837948

[Schlagi123]

If it is right to not having the protocol in the website tag, thats ok for me, but then, the bug is by the wiki page. I have a problem with inconsistent behavioure between documentation and an editors. And the only documentation I find as dao user, is the wiki. Nobody how is not so deep in OSM would read a talk aboute that, when the wiki page is so clear. And then he struggle, when he see website-tags without protocol and the validators of the editors do not warn about that.

So, for me there is a bug in JOSM or in the OSM-Wiki.

[stoecker]

I gave up trying to fix or even contribute to the wiki. People simply change the way how tags are handled without ever checking with editor authors or application programmers whether that's even feasible.

So for many years we have the situation that JOSM follows the wiki when possible and useful and ignores it where not.

[Schlagi123]

I understand that this is frustrating, but if there is no point of truth in the community, then the project is nearly dead, because nobody can build on something. In the worst case, editors drift apart. And then data quality can not be guaranteed.

Ok, what we do with that website problem? Nothing?

[stoecker]

Replying to Schlagi123:

I understand that this is frustrating, but if there is no point of truth in the community, then the project is nearly dead, because nobody can build on something. In the worst case, editors drift apart. And then data quality can not be guaranteed.

Well, the wiki isn't a definition of the database, but only a description of it. It can be nor is totally accurate. The whole concept of OSM is that it is a open project and this also includes disagreement between the components. While on the first view somewhat strange I still believe that's actually a very positive aspect and not an issue which must be fixed (thought I agree it took me years to come to this understanding).

Ok, what we do with that website problem? Nothing?

If there are URLs which are missing http: or https: (i.e. more than the domain name) that should be fixed. Single domains should still be ok and allowed, as this is simply what we see in the reality also.

[Schlagi123]

I understand your points. Thanks for the replay. I have closed the issue.