Talk:Tag (metadata)

From Wikipedia, the free encyclopedia
Jump to: navigation, search

Microformats proposal, 2009[edit]

For a proposal to allow Tagging on Wikipedia, see Wikipedia:WikiProject Microformats#MediaWiki issues — Preceding unsigned comment added by Pigsonthewing (talkcontribs) 20:51, 7 February 2009 (UTC)

Favorite and favourite[edit]

I noticed a spelling error in the Hash Tag section. I changed the spelling 'favorite' (incorrect) to 'favourite' (correct) — Preceding unsigned comment added by Jordaniair (talkcontribs) 14:56, 28 November 2011 (UTC)

Facebook tags[edit]

I think that Facebook should be added to the "popular websites that use tags" section. Not only is facebook a large website, but it's photo tagging feature is the first of it's kind (as far as I know, but I could be wrong.). I propose adding the following entry (please, any input would be appreciated)

  • Facebook – A social networking website that allows users to tag their friends in photos. Tagged friends will have their name display if you hold the mouse over their image, allowing for easy viewing of pictures of a specfic person, as opposed to photos taken by a specific person.

DaRkAgE7[Talk] 06:54, 15 August 2008 (UTC)

This is an odd case since Facebook uses the term "tagging" in a different way from what tagging means in Flickr, Delicious, YouTube, etc. - Facebook's tags are nothing to do with classification, just about identifying the people in pictures. We should probably note it in the article somehow though. Dreamyshade (talk) 07:20, 21 September 2008 (UTC)

Who on earth said tags were, first and foremost, for classification ? —Preceding unsigned comment added by 79.89.112.181 (talk) 01:03, 27 April 2009 (UTC)

The word "tag" can have a variety of meanings, but the subject of this article is tags as pieces of metadata that classify pieces of data. Tags as currently implemented on Facebook are links to other Facebook pages: see Tag (Facebook). Biogeographist (talk) 17:41, 11 March 2017 (UTC)

Major revision on September 21[edit]

Hi everybody! I just went through and changed most of the article (I'd worked on it offline). I'm wary of making such major changes unilaterally, especially since I'm somewhat biased, so please discuss if I went too far astray on any point. One issue is that I didn't fully address the concerns in the "History" section above — this is tricky and I need to find more decent citations for the history of tagging. Dreamyshade (talk) 07:20, 21 September 2008 (UTC)

For what it's worth, you can find a March 2002 reference to me using tags at http://web.archive.org/web/20020525043925/http://muxway.org/ (note the link in the tags, eg "obsess" is to /index.cgi?tag=obsess, for example.) While keywords are not new, I believe that tagging is a larger concept than just assigning keywords to things, however - I feel that it also includes the retrieval of the set of used terms/keywords/whatever upon view of the items. Additionally, I am reasonably sure that I named this. JoshuaSchachter (talk) 06:01, 22 September 2008 (UTC)

Hierarchical tags[edit]

Wikipedia's categorization system is similar to a tagging system, except it also allows hierarchies, in which tags are given tags. This should be mentioned. 71.167.63.236 (talk) 15:02, 20 October 2008 (UTC)

No matter whether those are tags or categories or whatever, we probably don't want to mention Wikipedia in the article; see Wikipedia:Self-references to avoid. Dreamyshade (talk) 00:35, 21 October 2008 (UTC)

Increasingly, tagging systems allow hierarchies and/or create hierarchies - thus in many Web systems the tags have become tied to specific indexes - in fact the hierarchical tagging of Wikipedia is part of what has led to this change. I would advocate removing "non-hierarchical" from the definition of tag -- but didn't want to make the change directly since I know it could be controversial. See for example PubMed's use of XML tagging, or other large-scale info systems (as well as wikipedia itself) for examples of hierarchical tags. — Preceding unsigned comment added by JAHendler (talkcontribs) 23:51, 24 June 2013 (UTC)

Hash tags[edit]

That subsection is not much use, since it tells us that Twitter (etc) messages can use hashtags (space in the subsection title; no space in the text: why?) but doesn't explain whether or how they differ from other types of tag. Loganberry (Talk) 01:14, 20 April 2009 (UTC)

  • Something on the origin of hashtags would also be good. A potential reference for expansion (if it's seen as reliable) would be the Twitter help page. Mike Peel (talk) 20:59, 22 January 2010 (UTC)
    • It would indeed. Reading currently it appears that hash tags did not exist before twitter, and of course Google is no help (it just finds twitter as well). Thanks, ♫ SqueakBox talk contribs 15:23, 8 April 2010 (UTC)
    • Current Wikipedia coverage on hashtags in this article is helpful and good. However, there's more to say. Hashtags could justifiably be broken out into its own article after more content is added. FYI, I'm currently using a hashtag to mark an #issue for resolution in MS Word documents that I've been writing and/or reviewing. Itohacs 13:26, 14 February 2011 (UTC) — Preceding unsigned comment added by Itohacs (talkcontribs)

Expanded tags[edit]

This text was added to the article today:

An expanded tag is a user-generated tag that captures more than conventional keywords, they describe some additional aspect of a data object. "Jumper 2.0 Tags the Enterprise". John Udell. 17 April 2009.  Text "Amazines" ignored (help); Check date values in: |date= (help) The expanded tag field takes the basic premise of tagging beyond just web pages to capture knowledge about structured and semi-structured data. It was pioneered by the Jumper 2.0 platform which collects expanded tags and stores them in a tag profile."Jumper Networks Press Release Jumper 2.0 Released under the GPL" (PDF). Jumper Networks, Inc. 26 March 2009.  Check date values in: |date= (help) The expanded tag fields allow users to input more than keyword knowledge, they can categorize the data, provide definitions, descriptions, annotations, add comments and notes to the data, capture the system metadata, ontologies and taxonomies, and much more.

It's only cited with two press releases, which is probably not enough to justify putting it in the article. It should have citations in reliable sources to show that it's not just a usage coined by one company. There has also been a lot of wiki activity supporting information about Jumper 2.0 recently (see the histories of Jumper 2.0 and Enterprise bookmarking), and I suspect conflict of interest, but I haven't been able to figure that out yet. Dreamyshade (talk) 18:13, 18 May 2009 (UTC)

I have expanded this with two other references one from the Journal of Information Science and one from the Business Intelligence Journal. Metadata is "data about data" it is any information that further defines another object of information. From the research it is clearly evident that bookmarking tags are both "tags" and "metadata". With the enormous popularity of bookmarking it seems relevant to the subject. —Preceding unsigned comment added by Rabbasher (talkcontribs) 02:33, 19 May 2009 (UTC)
My concern is that the phrase "expanded tag" is not in common use, and none of the three citations for that paragraph mention that phrase, as far as I can tell. In other words, both the Journal of Information Science and Business Intelligence Journal articles are good references for this topic, but they don't serve as verification for the the paragraph that you added. Non-notable, unverified neologisms don't belong in the article. Still, I agree that tags on bookmarks are relevant to this article; the "History and context" section mentions bookmarking. (Note that these comments are not about my removal of the "see also" link to enterprise bookmarking — I believe that link is not very important to this article, which is why I removed it, but I don't think it hurts the article so I don't mind leaving it for now.) Also, are you connected to Jumper 2.0 in any way? I have a potential conflict of interest because I work for Delicious, so I feel comfortable editing this article only because I disclose my affiliations on my user page and have read the relevant Wikipedia guidelines. Dreamyshade (talk) 03:08, 19 May 2009 (UTC)
Removed it for now, pending discussion and per the points above. Prodego talk 03:17, 19 May 2009 (UTC)

Triple tags[edit]

I've updated a broken link in this section. However, as far as I can tell, none of the links demonstrate the history of triple tags as described in this section - they don't disprove it either, they're just several isolated posts talking about tags of a similar form, making no reference to one another. There's no news-style discussion of "wider acceptance" or the development of the format. It's all primary sources, which is a bit awkward. -- Shimmin Beg (talk) 09:22, 5 August 2013 (UTC)

Better history?[edit]

It might be possible to research a better history of tags. Some programming text editors such as Emacs were extended to support tags; these tags were indices into software that could comprise more than one file, originally supporting operations on those tags for navigating to definitions and references of variables. I'm not sure when such early and limited tag facilities were invented; possibly in the early 1970s. David Spector (talk) 13:54, 30 October 2016 (UTC)

@David spector: That sounds interesting. If you can find a reliable source on that topic, please add a sentence or two and reference into the history section, which I revised today. There would also need to be some rationale for distinguishing the functionality you described from hyperlinks (that is, the source would have to clearly show how what you described is tagging, and not a kind of proto-hyperlinking). Thanks, Biogeographist (talk) 03:05, 11 March 2017 (UTC)

Well, Biogeographist, I think the Emacs use of tagging was indeed an early kind of hyperlinking (that doesn't exclude it from a history of tags, since it was one of the first uses of tags and was called tags). It did not have the facility to click a link embedded in program source text, but it did have the ability to go to the definition or reference of a tag in the source files, and such tags were automatically generated (although the author had to run a special tool to create or update the tag lists).

This early use of tagging was (and probably still is) also available in some program development systems, too (IDEs or PDSs). I believe that hyperlinks are different from both uses of tagging. The two uses of tagging are: first, as summaries of key contents of mail messages, websites, and other data, and second, as automatically-generated lists of definitions and references in editors, IDEs, and WYSIWYG formatted editors like Microsoft Word used to view items like variable names, function names, and index or table of contents entries or referenced document pages, having the functionality of instant source viewing. (The tags facility in Word is called cross-references.) I'm sure that the concise description in this paragraph could be improved and made clearer.

As to finding a reference to this Emacs facility, page http://www.gnu.org/software/emacs/manual/html_node/emacs/Tags.html is a good reference because it is authoritative due to being part of the programming manual for Emacs. Reference https://support.microsoft.com/en-us/help/284927/description-of-the-smart-tags-in-word describes Smart Tags in Word, which are words underlined with dotted lines that could be clicked--these are like automatically generated hyperlinks, not true tags. Reference https://support.office.com/en-us/article/Create-a-cross-reference-300b208c-e45a-487a-880b-a02767d9774b describes cross-references in Word, which are true tags.

If you agree that some of this information belongs in the history of tagging, please do the editing for me. I'm a bit fragile due to my current cancer treatment. Thanks, David Spector (talk) 12:29, 11 March 2017 (UTC)

@David spector: Thanks very much for those interesting references, and please take care of yourself. I am still looking at the Emacs references, which I think may be relevant and historically significant. The Microsoft Word references, on the other hand, are not directly relevant: Microsoft's smart tags already have their own article, where they are described as a combination of named-entity recognition and selection-based search, and Word's cross-references are not tags; they are, well, cross-references. At a very high level of abstraction (as in graph theory) these may (or may not) all be the same thing (I don't know—that would require more analysis than I have done), but the definition of tag in this article is not at that high level of abstraction. In this article, a tag is defined as a piece of metadata attached to another piece of data that classifies that data; it is the classification function of the tag that is most important, it seems to me.
Having said all that, I am now thinking that Emacs "tags" may be cross-references (but I still haven't finishing investigating), but they may merit mentioning in the history section for their historical significance and to differentiate them from the kind of tags described in this article. The relationship between all of these concepts is extremely interesting. For example, the article on named-entity recognition (NER) (which as I mentioned above is one aspect of Microsoft's smart tags) mentions NER systems "that use linguistic grammar-based techniques as well as statistical models, i.e. machine learning"; on the one hand, the metadata tags defined in this article could be used as the basis for grammar-based NER, and on the other hand, more sophisticated NER using machine-learning techniques could extract a grammar from a large corpus of user-generated tags. Thanks again, Biogeographist (talk) 14:12, 11 March 2017 (UTC)

I agree with your thoughts, specifically that Emacs tags are not "tags" in the sense of keywords that classify data into orthogonal categories. However, in history, that's exactly what the designers of Emacs chose to call their generated and hyperlinked cross-references. And yes, depending on the sophistication and discipline with which humans choose tags, those tags can be used to generate context-free grammars suitable for AI use (for example, to encode expert knowledge). Keep going with your analysis, and then edit the history; I'm certain you will do a great job. David Spector (talk) 15:57, 11 March 2017 (UTC)

Merge proposal, March 2017[edit]

I propose that the article on tags assigned to files that was created yesterday by User:Equalhuman should be merged into this article. The bold term in the lead sentence of tags assigned to files is tag, and the lead sentence of that article repeats nearly verbatim the lead sentence of this article. The content of that article is brief enough that it can easily be merged into this article. The only article that links to that article is this article. In summary, I see no good reason why tags assigned to files deserves to be an article separate from this one. Biogeographist (talk) 19:54, 8 March 2017 (UTC)

I can agree to your critique. tags assigned to files is not perfectly perfect yet. However, I think, that it might answer different question:
Tag (metadata) answers the general question "what is a tag" and therefore gives a precise and universal definition of the term and the general principles. tags assigned to files answer the question for someone who wants to tag files and therefore would like to know possible solutions for his or her desire.
I thought merging both answers would result in making Tag (metadata) less universal and therefore spoiling the definition audience and tags assigned to files more background heavy and therefore covering the answer in the general information and making it more unsatisfying for the "solution needed" audience. Therefore I choose to make a new article tags assigned to files. I just wanted to explain my afterthoughts. I can't answer if it's right. Thank you for investing time to preserve wikipedia and asking this legitimate question. Equalhuman (talk) 10:03, 10 March 2017 (UTC)
@Equalhuman: If your intention was to write an article with a detailed explanation of how to tag files using various software systems, that might not be an appropriate subject for a Wikipedia article per WP:NOTHOWTO; there are many other websites where you can post that kind of information, such as the Stack Exchange network of question-and-answer websites, which already has many good pages on the subject of tagging files. If your intention was to point out that file tagging exists, and to list some of the major systems for file tagging (which, by the way, should include the cross-platform TagSpaces application, as well as the Tagsistant file system for Linux), that is an appropriate subject that can easily be included in Tag (metadata). I have added the section Assigned to computer files; please look at it and tell me what you think, and add any information that you think is missing. I think that section constitutes a merging of Tags assigned to files into this article, and if you agree, then I will redirect that article to this one, finishing the merge. Thanks very much for your work, Biogeographist (talk) 02:25, 11 March 2017 (UTC)

  YesY Merger complete. — Merge completed; no objections after more than a week. Biogeographist (talk) 11:05, 24 March 2017 (UTC)