Ad-hoc, mix-and-match tag hierarchies?

Ad-hoc, mix-and-match tag hierarchies?

[Lawrence Bottorff]

[-- Attachment #1: Type: text/plain, Size: 885 bytes --]

Is it possible to have two or more tags that are "peers," i.e., all equal,
not in a hierarchy, be in an ad-hoc, as-needed way be hierarchical? For
example, I have the tags *org-mode, lisp, *and *emacs, *and I want to have
a header with the tags

* my header    :emacs:org-mode:lisp:

So the above should be an ad-hoc hierarchy of

- emacs
   - org-mode
      - lisp

As I understand, the order indicates the level in a hierarchy. True? So the
above has *emacs* at the top of the tag hierarchy, then *org-mode, *then *lisp.
*Correct? So yes, I could simply set up this hierarchy. But what I really
want is to not have these tags in any set hierarchy, rather, be able to use
them independently, mix-and-match, e.g.,

* Another Header    :emacs:lisp:

or maybe

* Yet Another Header   :lisp:

and this would be just about non-Emacs, non-org-mode Lisp. Is this
mix-and-match possible?

LB

[-- Attachment #2: Type: text/html, Size: 1192 bytes --]

Re: Ad-hoc, mix-and-match tag hierarchies?

[John Kitchin]

[-- Attachment #1: Type: text/plain, Size: 1733 bytes --]

I was not aware of any implied hierarchy with tags. The order is not
important as far as I know, and what you describe as mix-and-match seems
ok. There is an idea of inheritance, e.g. sub-headings can inherit tags
from higher headings.

In any case, you should be able to use agenda queries to find headings that
are lisp and not emacs, lisp+orgmode, etc.

I am not sure about the org-mode tag. In the past, emacs splits that into
:org:mode: for me.

What do you want to do with the tags?

John

-----------------------------------
Professor John Kitchin
Doherty Hall A207F
Department of Chemical Engineering
Carnegie Mellon University
Pittsburgh, PA 15213
412-268-7803
@johnkitchin
http://kitchingroup.cheme.cmu.edu


On Thu, Feb 1, 2018 at 7:07 AM, Lawrence Bottorff <borgauf@gmail.com> wrote:

> Is it possible to have two or more tags that are "peers," i.e., all equal,
> not in a hierarchy, be in an ad-hoc, as-needed way be hierarchical? For
> example, I have the tags *org-mode, lisp, *and *emacs, *and I want to
> have a header with the tags
>
> * my header    :emacs:org-mode:lisp:
>
> So the above should be an ad-hoc hierarchy of
>
> - emacs
>    - org-mode
>       - lisp
>
> As I understand, the order indicates the level in a hierarchy. True? So
> the above has *emacs* at the top of the tag hierarchy, then *org-mode, *then
> *lisp. *Correct? So yes, I could simply set up this hierarchy. But what I
> really want is to not have these tags in any set hierarchy, rather, be able
> to use them independently, mix-and-match, e.g.,
>
> * Another Header    :emacs:lisp:
>
> or maybe
>
> * Yet Another Header   :lisp:
>
> and this would be just about non-Emacs, non-org-mode Lisp. Is this
> mix-and-match possible?
>
> LB
>

[-- Attachment #2: Type: text/html, Size: 2686 bytes --]

Re: Ad-hoc, mix-and-match tag hierarchies?

[Kaushal Modi]

[-- Attachment #1: Type: text/plain, Size: 848 bytes --]

On Thu, Feb 1, 2018 at 10:30 AM John Kitchin <jkitchin@andrew.cmu.edu>
wrote:

> I was not aware of any implied hierarchy with tags. The order is not
> important as far as I know, and what you describe as mix-and-match seems
> ok. There is an idea of inheritance, e.g. sub-headings can inherit tags
> from higher headings.
>

Same here.


> I am not sure about the org-mode tag.
>

"-" is an invalid tag character.. the tag can be "org_mode", but not
"org-mode". Last time I tried using hyphen, I remember that the tag
auto-alignment would stop working (C-u C-c C-q) and also the tag face won't
be applied.

@Lawrence:

You can have tag inheritance this way:

* Emacs :emacs:
Emacs stuff
** Lisp :lisp:
Emacs lisp stuff here
*** Org :org_mode:
Emacs + Lisp + Org mode stuff
** Org :org_mode:
Emacs + Org mode stuff (but not lisp)
-- 

Kaushal Modi

[-- Attachment #2: Type: text/html, Size: 1584 bytes --]

Re: Ad-hoc, mix-and-match tag hierarchies?

[Lawrence Bottorff]

[-- Attachment #1: Type: text/plain, Size: 2642 bytes --]

So a left-to-right listing of (colon-separated) tags after the heading
cannot imply a higher-to-lower hierarchical order? So there is no hierarchy
unless you create it, e.g.,

(setq org-tag-alist '((:startgrouptag)
                       ("GTD")
                      (:grouptags)
                      ("Control")
                      ("Persp") ...

-- is what I'm hearing.

What I meant to do is to be able to use tags alone or together. And when
used together, to somehow imply hierarchy. So if I have

* header 1            :emacs:

** header 1.1       :orgmode:

*** header 1.1.1   :lisp:
stuff about Emacs' org-mode's lisp code

this implies subject hierarchy simply from the headers' hierarchy. However
this

* header 1   :emacs:
stuff about Emacs
* header 2  :emacs:orgmode:lisp
stuff about Emacs' org-mode's lisp code
* header 3 :lisp:
stuff about just Common Lisp
* header 4 :emacs:orgmode:
stuff about Emacs' org-mode
* header 5  :emacs:
more stuff about Emacs

is all peers header-wise, while the tags (if using left-to-right listing to
mean higher to lower in hierarchy) tell us the depth level of
specialization of a topic. The first set of headers is relying (clumsily)
on the header level to indicate depth of specialization, which is not
always realistic. If I wanted, ad-hoc, to indicate levels of some hierarchy
with tags, that is, not be forced to create empty superior headers, it
seems  there's no prescribed way to do this. And creating tag hierarchies
by hand, i.e., (setq org-tag-alist '((:startgrouptag)..., makes me play the
combinatorics game of imagining all different possible hierarchy
combinations. Or am I missing something?

On Thu, Feb 1, 2018 at 10:36 AM, Kaushal Modi <kaushal.modi@gmail.com>
wrote:

> On Thu, Feb 1, 2018 at 10:30 AM John Kitchin <jkitchin@andrew.cmu.edu>
> wrote:
>
>> I was not aware of any implied hierarchy with tags. The order is not
>> important as far as I know, and what you describe as mix-and-match seems
>> ok. There is an idea of inheritance, e.g. sub-headings can inherit tags
>> from higher headings.
>>
>
> Same here.
>
>
>> I am not sure about the org-mode tag.
>>
>
> "-" is an invalid tag character.. the tag can be "org_mode", but not
> "org-mode". Last time I tried using hyphen, I remember that the tag
> auto-alignment would stop working (C-u C-c C-q) and also the tag face won't
> be applied.
>
> @Lawrence:
>
> You can have tag inheritance this way:
>
> * Emacs :emacs:
> Emacs stuff
> ** Lisp :lisp:
> Emacs lisp stuff here
> *** Org :org_mode:
> Emacs + Lisp + Org mode stuff
> ** Org :org_mode:
> Emacs + Org mode stuff (but not lisp)
> --
>
> Kaushal Modi
>

[-- Attachment #2: Type: text/html, Size: 4326 bytes --]

Re: Ad-hoc, mix-and-match tag hierarchies?

[John Kitchin]

I don't think you are missing anything. Your first example leverages inheritance (and you can turn that off if you don't want it), and your second example doesn't.

In any case, the order of the tags is irrelevant. To me the real question is what do you want to use them for?

I use them in agenda queries, and there the order is not important. That
is a nice feature, since you don't have to know the hierarchy, and can
query emacs+lisp or lisp+emacs for example and get the same things.
Suppose you have the headers at the end of this email.

These:

(org-tags-view nil "emacs+lisp")
(org-tags-view nil "lisp+emacs")

 both show in the agenda (these are in a file called 2018-02-03.org)

  2018-02-03: header 1.1.1                                                     :emacs:orgmode::lisp:
  2018-02-03: header b                                                          :emacs:orgmode:lisp:
  2018-02-03: header B                                                          :lisp:emacs:orgmode:

Note that header 1.1.1 has all the tags from inheritance, and the other
two have all the tags in different orders. You can use a query like
"emacs-orgmode" to get headlines tagged emacs and not tagged orgmode.







* header 1            :emacs:

** header 1.1       :orgmode:

*** header 1.1.1   :lisp:
stuff about Emacs' org-mode's lisp code

* header a                                                            :emacs:
stuff about Emacs
* header b  :emacs:orgmode:lisp:
stuff about Emacs' org-mode's lisp code
* header B :lisp:emacs:orgmode:
The tag order is not important

* header c                                                             :lisp:
stuff about just Common Lisp
* header d                                                    :emacs:orgmode:
stuff about Emacs' org-mode
* header e                                                            :emacs:
more stuff about Emacs


Lawrence Bottorff writes:

> So a left-to-right listing of (colon-separated) tags after the heading
> cannot imply a higher-to-lower hierarchical order? So there is no hierarchy
> unless you create it, e.g.,
>
> (setq org-tag-alist '((:startgrouptag)
>                        ("GTD")
>                       (:grouptags)
>                       ("Control")
>                       ("Persp") ...
>
> -- is what I'm hearing.
>
> What I meant to do is to be able to use tags alone or together. And when
> used together, to somehow imply hierarchy. So if I have
>
> * header 1            :emacs:
>
> ** header 1.1       :orgmode:
>
> *** header 1.1.1   :lisp:
> stuff about Emacs' org-mode's lisp code
>
> this implies subject hierarchy simply from the headers' hierarchy. However
> this
>
> * header 1   :emacs:
> stuff about Emacs
> * header 2  :emacs:orgmode:lisp
> stuff about Emacs' org-mode's lisp code
> * header 3 :lisp:
> stuff about just Common Lisp
> * header 4 :emacs:orgmode:
> stuff about Emacs' org-mode
> * header 5  :emacs:
> more stuff about Emacs
>
> is all peers header-wise, while the tags (if using left-to-right listing to
> mean higher to lower in hierarchy) tell us the depth level of
> specialization of a topic. The first set of headers is relying (clumsily)
> on the header level to indicate depth of specialization, which is not
> always realistic. If I wanted, ad-hoc, to indicate levels of some hierarchy
> with tags, that is, not be forced to create empty superior headers, it
> seems  there's no prescribed way to do this. And creating tag hierarchies
> by hand, i.e., (setq org-tag-alist '((:startgrouptag)..., makes me play the
> combinatorics game of imagining all different possible hierarchy
> combinations. Or am I missing something?
>
> On Thu, Feb 1, 2018 at 10:36 AM, Kaushal Modi <kaushal.modi@gmail.com>
> wrote:
>
>> On Thu, Feb 1, 2018 at 10:30 AM John Kitchin <jkitchin@andrew.cmu.edu>
>> wrote:
>>
>>> I was not aware of any implied hierarchy with tags. The order is not
>>> important as far as I know, and what you describe as mix-and-match seems
>>> ok. There is an idea of inheritance, e.g. sub-headings can inherit tags
>>> from higher headings.
>>>
>>
>> Same here.
>>
>>
>>> I am not sure about the org-mode tag.
>>>
>>
>> "-" is an invalid tag character.. the tag can be "org_mode", but not
>> "org-mode". Last time I tried using hyphen, I remember that the tag
>> auto-alignment would stop working (C-u C-c C-q) and also the tag face won't
>> be applied.
>>
>> @Lawrence:
>>
>> You can have tag inheritance this way:
>>
>> * Emacs :emacs:
>> Emacs stuff
>> ** Lisp :lisp:
>> Emacs lisp stuff here
>> *** Org :org_mode:
>> Emacs + Lisp + Org mode stuff
>> ** Org :org_mode:
>> Emacs + Org mode stuff (but not lisp)
>> --
>>
>> Kaushal Modi
>>


--
Professor John Kitchin
Doherty Hall A207F
Department of Chemical Engineering
Carnegie Mellon University
Pittsburgh, PA 15213
412-268-7803
@johnkitchin
http://kitchingroup.cheme.cmu.edu