Possible Duplicates:
Another kind of tag hierarchy/relationship
Could we make tags imply other tags?

i already suggested this here but i think this suggestion deserves it's own "question".

this suggestion tries to solve the problem highlighted here

synonyms are an answer to different tags which mean the same ([oop] and [object-oriented]), but there tags which are strongly related but do not mean the same ([php5] to [php]). this is where tag inheritance comes in.

here is a copy of my "answer" to the php5 "question":

i suggest tag inheritance.

something like php5 extends php and php4 extends php

this means everything tagged as [php5] is automatically also tagged [php], but a question tagged as [php] will not automatically get tagged as [php5].

people interested only in php5 questions can follow only the [php5] tag. people interested in php in generall need to follow only [php] and will automatically also follow [php5] and [php4]

having [php] and [php5] as synonyms will not allow this differentiation.

note: tag inheritance will not solve the problem of people tagging their question wrong ([php5] when it is not php5 specific), that can only be solved by educating and retagging.

toying around with the tag inheritance idea a bit further, using python as example:

python2x extends python
python3x extends python

python-old extends python
python-current extends python
python-future extends python

currently python 2.7 and python 3.1 are both stable releases.

python26 extends python2x, python-old
python27 extends python2x, python-current
python30 extends python3x, python-old
python31 extends python3x, python-current
python32 extends python3x, python-future

eventually python 2.x will be declared dead, python 3.2 will be released, and maybe they will start working on python 4. then the inheritance tree can be changed to look like this.

python26 extends python2x, python-old
python27 extends python2x, python-old
python30 extends python3x, python-old
python31 extends python3x, python-old
python32 extends python3x, python-current
python33 extends python3x, python-future
python40 extends python4x, python-future

this will allow following scenarios:

  • searching for [python] will find all question tagged with any of the python tags
  • searching for [python3x] will find [python30], [python31] and [python32]
  • searching for [python3x] and [python-current] will find only [python31] (in the future it will find [python32])
  • searching for [python-current] will find [python27] and [python31] (in the future it will find [python32])
  • searching for [python-old] will find [python26] and [python30]
  • searching for [python-old] and [python2x] will find only [python26] (in the future it will find [python26] and [python27])

there is nothing principally new in tag inheritance. tag inheritance can be done today when everyone (EVERYONE) who tags a specific tag ([php5]) always (ALWAYS) also tags the more general tag ([php]). the problem is people don't do that.

the advantage of tag inheritance is that it is done automatically. furthermore, like the python example shows, we can do easy mass retagging by changing the inheritance tree.

marked as duplicate by dmckee, waiwai933, Ether, random Aug 12 '10 at 0:52

This question has been asked before and already has an answer. If those answers do not fully address your question, please ask a new question.

  • 1
    This is roughly a duplicate of Another kind of tag hierarchy/relationship – dmckee Aug 11 '10 at 15:54
  • Please don't introduce the diamond problem into the tag system ;) – kennytm Aug 11 '10 at 15:55
  • i do not agree that this is an "exact duplicate". both alleged duplicates suggest a tag hierarchy like a tag tree, which jeff refuses to implement, which i wholeheartedly agree to refuse to implement because of the inherent problem of tag trees. i did not suggest tag trees, i suggested tag inheritance. tag inheritance does not have the inherent problem of tag trees because of the multiple inheritance. while i can accept a decision to not implement my suggestion, i do not accept that my suggestion is an "exact duplicate" of the other suggestions. – lesmana Aug 12 '10 at 10:22
  • related: implicit tagging hierarchy. – Tobias Kienzler Aug 12 '10 at 12:04
  • IMHO, there are valid reasons for and against the tag inheritance. Could creating the links (actually a DAG) without enforcing it be the solution? I mean to offer the implied tags to the user with a checkbox for each and let them decide. There could be even two kinds of links: strong and weak where the former/latter would offer a checked/unchecked box, so minimizing the work the user had to do to get it right. – maaartinus Jan 28 '11 at 6:40

It could also have the benefit of keeping the higher up tag 'under the hood'. You wouldn't need to see PHP, just PHP5, but it could keep track of it as if the tag were there.

Not the answer you're looking for? Browse other questions tagged .