29
30
Fork 85

Potential name-clash between country tag and other two-letter tags #965

Offen
2019-05-28 22:37:03 +00:00 von jzarl geöffnet · 15 Kommentare
Mitglied

Looking through the list of two-letter tags, I noticed that mk is used to reference to Matthias. This may become a problem at some point when we have some article/event/news-item related to the country of North Macedonia.

Should we proactively do something about this, or do we wait until this is more than a theoretical issue?

On a related note, we also have the two-letter tag ga for general assembly, but that is even less of a problem: ga is not currently an assigned country code. The same applies to os.

Looking through the list of two-letter tags, I noticed that `mk` is used to reference to Matthias. This may become a problem at some point when we have some article/event/news-item related to the country of North Macedonia. Should we proactively do something about this, or do we wait until this is more than a theoretical issue? On a related note, we also have the two-letter tag `ga` for general assembly, but that is even less of a problem: ga is not currently an assigned country code. The same applies to `os`.
jzarl hat das Label
tagging
2019-05-28 22:37:14 +00:00 hinzugefügt
Autor
Mitglied

I've missed the aa tag (currently used for "Auswärtiges Amt"). I guess we can remove that one completely, though...

I've missed the `aa` tag (currently used for "Auswärtiges Amt"). I guess we can remove that one completely, though...
jzarl hat den Titel von Potential name-clash between country tag and username zu Potential name-clash between country tag and other two-letter tags 2019-05-28 22:53:01 +00:00 geändert
Mitglied

Thank you so much @jzarl for your great cleanup work!

I think that just as we define rules for tags relating to countries, we should also have rules how tags relating to people should be built. It seems like first_lastname is often used, other possibilities would be first-lastname, first.lastname or even just lastname (a dangerous proposal for somebody named Müller, though). The login is probably not a good option because of the potential name clashes with country codes.

Thank you so much @jzarl for your great cleanup work! I think that just as we define rules for tags relating to countries, we should also have rules how tags relating to people should be built. It seems like first_lastname is often used, other possibilities would be first-lastname, first.lastname or even just lastname (a dangerous proposal for somebody named Müller, though). The login is probably not a good option because of the potential name clashes with country codes.
Besitzer

I even doubt that we need tags for people working with us:

  1. Matthias is probably involved in 80% of the publications of the last years. Shall we also tag all of them?
  2. We haven't done the same for most other people who visited events and gave talks.

So either we apply the according tags to all other people who gave talk, or we leave it out completely. I would suggest the second option and get rid of these old, rather useless tags for people.

I even doubt that we need tags for people working with us: 1. Matthias is probably involved in 80% of the publications of the last years. Shall we also tag all of them? 2. We haven't done the same for most other people who visited events and gave talks. So either we apply the according tags to all other people who gave talk, or we leave it out completely. I would suggest the second option and get rid of these old, rather useless tags for people.
Mitglied

Actually I think that, for example, recent news and recent/future events tagged "max.mehl" could be a nice addition to https://fsfe.org/about/mehl/ - just like events and news tagged "de" are listed on https://fsfe.org/de/.

Actually I think that, for example, recent news and recent/future events tagged "max.mehl" could be a nice addition to https://fsfe.org/about/mehl/ - just like events and news tagged "de" are listed on https://fsfe.org/de/.
Autor
Mitglied

i think just removing all tags that are not used correctly is not a good solution. OTOH, https://fsfe.org/tags/tagged-matthias_kirschner.en.html is rather useless as it is.

So, I propose removing mk from the newsletters, but not from the events and the articles that he has written.
Btw. how is the author tagging implemnted? campaigns/generalpurposecomputing/secure-boot-analysis.en.xhtml can be found using the tags page, but there is no tags section in the article, just the author id, which is not mk, but kirschner...

Similarly, other people should IMO be tagged when they are speaking at some event or if they are author of an article.

i think just removing all tags that are not used correctly is not a good solution. OTOH, https://fsfe.org/tags/tagged-matthias_kirschner.en.html is rather useless as it is. So, I propose removing mk from the newsletters, but not from the events and the articles that he has written. Btw. how is the author tagging implemnted? `campaigns/generalpurposecomputing/secure-boot-analysis.en.xhtml` can be found using the tags page, but there is no tags section in the article, just the author id, which is not `mk`, but `kirschner`... Similarly, other people should IMO be tagged when they are speaking at some event or if they are author of an article.
Mitglied

There is no author tagging, tagging is only for news and events. The secure boot analysis is listed on the tag page because of news/2012/news-20120601-01.en.xml.

There is no author tagging, tagging is only for news and events. The secure boot analysis is listed on the tag page because of `news/2012/news-20120601-01.en.xml`.
Autor
Mitglied

Ah, should have thought of that... Do you think it possible to feed the author information from articles into the tagging system?

So that leaves tagging people who are speaking at some event. For now, I guess that removing the tag for Matthias from the news entries only seems like a good middle ground. For the future, it would be nice to have people tagged in events for which they are relevant, but I'm unsure how we could enforce/encourage this.

Ah, should have thought of that... Do you think it possible to feed the author information from articles into the tagging system? So that leaves tagging people who are speaking at some event. For now, I guess that removing the tag for Matthias from the news entries only seems like a good middle ground. For the future, it would be nice to have people tagged in events for which they are relevant, but I'm unsure how we could enforce/encourage this.
Mitglied

Tagging is currently only implemented for news and events. It might be possible to extend that to other articles, but it would probably require quite some work in the underlying scripts.

Tagging is currently only implemented for news and events. It might be possible to extend that to other articles, but it would probably require quite some work in the underlying scripts.
Autor
Mitglied

I've created a new issue for the automatic tagging of articles: #986. Maybe we can pick this idea up some time in the future...

I've created a new issue for the automatic tagging of articles: #986. Maybe we can pick this idea up some time in the future...
Mitglied

Returning to the original issue: could we agree that 2-letter-codes should be reserved for countries, and that we want to change all other tags into a more-than-2-letter form? Are there actually any 2-letter tags left which aren't countries?

Returning to the original issue: could we agree that 2-letter-codes should be reserved for countries, and that we want to change all other tags into a more-than-2-letter form? Are there actually any 2-letter tags left which aren't countries?
Autor
Mitglied

I think we do agree on this. But more important in this case would be agreement of the content creators.

Regarding your second question: yes, there are still non-country two-letter tags. A list of all two-letter tags generated from the tags/.tags.en.xml file:

aa;aa;
at;Austria;
ch;Schweiz;
cz;cz;
de;Germany;
dk;dk;
ee;ee;
el;el;
es;Spain;
eu;European Union;
fi;Finland;
fr;France;
ga;General Assembly 2017;
gb;Great Britain;
it;Italy;
nl;Dutch;
os;os;
pt;pt;
ro;ro;
ru;ru;
se;Sweden;
sk;sk;
tr;tr;
un;un;
al;Albania;
at;Austria;
be;Belgium;
bg;Bulgaria;
ch;Schweiz;
cz;cz;
de;Germany;
ee;ee;
en;en;
es;Spain;
eu;European Union;
fi;Finland;
fr;France;
gb;Great Britain;
gr;Greece;
hr;Croatia;
hu;hu;
it;Italy;
jp;Japan;
mk;Matthias Kirschner;
nl;Dutch;
no;no;
pa;pa;
pl;Poland;
pt;pt;
ro;ro;
se;Sweden;
sk;sk;
tr;tr;

I still see the two previously mentioned examples "aa" and "mk".
Btw. it seems to me that "nl" is still at least partially used as a language tag. Otherwise, why would the tag be named "Dutch" on some pages?

I *think* we do agree on this. But more important in this case would be agreement of the content creators. Regarding your second question: yes, there are still non-country two-letter tags. A list of all two-letter tags generated from the `tags/.tags.en.xml` file: ``` aa;aa; at;Austria; ch;Schweiz; cz;cz; de;Germany; dk;dk; ee;ee; el;el; es;Spain; eu;European Union; fi;Finland; fr;France; ga;General Assembly 2017; gb;Great Britain; it;Italy; nl;Dutch; os;os; pt;pt; ro;ro; ru;ru; se;Sweden; sk;sk; tr;tr; un;un; al;Albania; at;Austria; be;Belgium; bg;Bulgaria; ch;Schweiz; cz;cz; de;Germany; ee;ee; en;en; es;Spain; eu;European Union; fi;Finland; fr;France; gb;Great Britain; gr;Greece; hr;Croatia; hu;hu; it;Italy; jp;Japan; mk;Matthias Kirschner; nl;Dutch; no;no; pa;pa; pl;Poland; pt;pt; ro;ro; se;Sweden; sk;sk; tr;tr; ``` I still see the two previously mentioned examples "aa" and "mk". Btw. it seems to me that "nl" is still at least partially used as a language tag. Otherwise, why would the tag be named "Dutch" on some pages?
Autor
Mitglied

Further inspection of the "Dutch" label for the "nl" tag reveals that this is probably more an incorrect translation rather than an incorrect tag:

$ grep -l 'Dutch\s*</tag>'  $(tools/tagtool/tagtool.sh --find-tags nl )
news/2017/news-20170302-01.en.xhtml
news/2017/news-20170302-01.nl.xhtml
Further inspection of the "Dutch" label for the "nl" tag reveals that this is probably more an incorrect translation rather than an incorrect tag: ``` $ grep -l 'Dutch\s*</tag>' $(tools/tagtool/tagtool.sh --find-tags nl ) news/2017/news-20170302-01.en.xhtml news/2017/news-20170302-01.nl.xhtml ```
Autor
Mitglied

On a side-note: if anyone has a translated list of country codes, we could add the tag labels automatically via a script...

On a side-note: if anyone has a translated list of country codes, we could add the tag labels automatically via a script...
Mitglied

That might be dangerous, since we might add the country-code label to 2-letter-tags that were actually meant to mean something else than the country.

That might be dangerous, since we might add the country-code label to 2-letter-tags that were actually meant to mean something else than the country.
Besitzer

I have removed aa meanwhile.

From your list, problematic tags might be os (for openstandards), mk (Matthias Kirschner) and nl (Dutch + Newsletter).

For mk, this becomes problematic if we use the FSFE username in the introduction of #986.

I have removed `aa` meanwhile. From your list, problematic tags might be `os` (for openstandards), `mk` (Matthias Kirschner) and `nl` (Dutch + Newsletter). For `mk`, this becomes problematic if we use the FSFE username in the introduction of #986.
Anmelden, um an der Diskussion teilzunehmen.
Kein Meilenstein
Niemand zuständig
3 Beteiligte
Nachrichten
Fällig am
Das Fälligkeitsdatum ist ungültig oder außerhalb des zulässigen Bereichs. Bitte verwende das Format „jjjj-mm-tt“.

Kein Fälligkeitsdatum gesetzt.

Abhängigkeiten

Keine Abhängigkeiten gesetzt.

Referenz: FSFE/fsfe-website#965
Keine Beschreibung angegeben.