Instruments with disambiguation comments

I noticed that quite a few of the newly added instruments have disambiguation comments that mostly duplicate the description blurb. Since the description is shown in the search results when adding instrument credits I think those search results are becoming a little too cluttered (see image). They are also untranslatable of course, which doesn’t help. Finally, some are a bit long. :wink:

Just a heads up. :slight_smile:

Slightly unrelated: I suspect we’ll eventually just remove the descriptions, because most of them are uncredited copy-pastes from Wikipedia.

Yes, I’ve read about that. But as long as those descriptions are very short (one short sentence), they can be useful in search results. Wikipedia often starts with an entire paragraph, so that would be way too long, and besides, not every instrument is on Wikipedia in the first place.

Basically this: they are added as I see fit, to be disambiguations (explicitly these Maori instruments, as I felt they where confusable with eachother somewhat.)

The “Description” section is a remnant of when instruments where just text-entries, nested in a tree, not entities with mbids in their own right. Another remnant is the “child of” relationship which is being phased out (I’m working on this) as well as things like INST-330

We discussed this in #mb, solutions to this include moving description text to annotations (and adding annotations to instruments if they do not have them yet), removing description text from searchview and just dropping descriptions altogether (because it mostly duplicates Wikipedia text without a credit)

Generally I try to make the disambiguations as short but descriptive as possible, and I reword the descriptions, explicitly, to make them not just be copypasta from wikipedia.

I’ve just started working on instrument requests, and I’ve just started adding and working on these instruments. I will generally add the text I feel is sufficient - as well as the links and rels possible, to identify and connect new (and old) instruments. It’s kind of a huge endeavour (as many as 200 open instrument requests!) in the future I might go back on these I’ve added and redefine or remove text (if at all necessary then, see previous paragraphs about solutions) but I will not do so, now.

FTR, there are also plenty of instruments without a disambiguation or a description (I’m working on that, too)

4 Likes

A short time ago, I was about to file an issue to add the ability to translate instruments description/disambiguation (?) field.

Excerpts from Wikipedia definitely does not fit the needs of the relationship editor.

I appreciate your efforts, but this sounds a bit arbitrary. And why not use the description to tell similar instruments apart? Pros: the description is shown in the search results, the description can be translated, and Wikipedia can’t replace these descriptions anyway. Cons: the description field might one day be dropped, but that has been the case for ages, and it still hasn’t happened.

Neither of these “solutions” are improvements, unless there is some reason that de instrument description field has to die. Annotations and disambiguation comments can’t be translated and that won’t be possible for a long time (if we even want that). Instrument descriptions already exist. Please note that this criticism isn’t aimed at you (you just happened to bring it up), but I’ve heard this a couple of times now without ever seeing a reason for this move.

I really appreciate your hard work. Clearing the instrument backlog is a huge task and all the interrelations make it a complex one too. But I’m a translator too so translatability is rather high on my wish list.

4 Likes

[quote=“mfmeulenbelt, post:6, topic:105752”]
Neither of these “solutions” are improvements, unless there is some reason that de instrument description field has to die.[/quote]

Most of the descriptions are copyright violations. So yes, it has to die. (The ticket is
MBS-8764.)

mfmeulenbelt: it seems to me that you don’t see that description-field does what that the disambiguation-field was meant for as erroneous, but right, and that the reason is that (supposedly) descriptions are translatable, but disambigs are not. correct?

Then the right thing to do is find a way to make disambiguations translatable (for more than just instruments, surly too.) as it is the object of the disambig, to disambiguate, it is the object of the description to describe.

[quote=“chirlu, post:7, topic:105752”]
Most of the descriptions are copyright violations. [/quote]
I explicitly reword them to avoid this, as I’ve said, and I’d be happy about adding them as annotations instead (again, reworded, so as to make sure there are no copyright problems. ) (this is also why I don’t just copy “disambiguation comments” from tickets verbatim as well.)

~Cat

2 Likes

Well, I’m not really looking forward to translating a few thousand “member of” disambiguation comments for artists (and its variations). Really the only place where I’d like these things translatable is with instruments, which makes that description field so convenient. And even if at this very moment we decide that translatable disambiguation comments are desirable, with the current development speed we can expect that to happen somewhere in the 2020s.

You’re right that I want the description field to do what disambiguation comments were really meant for, but the reason for that is that the disambiguation comment in its current incarnation is inadequate.

I’m not sure if instruments will have to be described in much more detail than what is possible in one sentence. We have relationships for their inventor, related or derived instruments and traditional area. More can be added if necessary, and if we really want to be exhaustive, there’s Wikipedia.

The contents of the field have nothing to do with existence of the field itself. @CatQuest is already rewording those descriptions anyway, whether they end up in the disambiguation comment or the description.

I think I’ll leave it at this. I’ve given my arguments and reasons. I can only ask you to consider them fairly and not dismiss them out of hand.

Also, I would really like to help with all this, like writing short Wikipedia articles for instruments that don’t have one yet (as long as I can find credible sources of course). In that case we wouldn’t need too much information in the description or annotation or wherever.

3 Likes

[quote=“mfmeulenbelt, post:9, topic:105752”]
The contents of the field have nothing to do with existence of the field itself. @CatQuest is already rewording those descriptions anyway, whether they end up in the disambiguation comment or the description.[/quote]

Which doesn’t help at all, because an uncredited modified copy is still a copyright violation. It’s true that this is technically not the description field’s fault, but a field that wants to duplicate what is already on Wikipedia just invites copy-and-paste action.

I still don’t understand what issue you have with showing (the start of) the Wikipedia article where the description is shown now. It would have a translation (from Wikipedia) automatically, and if not, anyone could add a missing Wikipedia article in the relevant language.

3 Likes

while I agree with you, I would like to make it clear that fact is not copyrightable, and my words are my own, I am very much writing things myself, not just “modifying” copypasta. I also take information for my “description” (and disambiguation, for that matter) not just from english wikipedia, but other lang. wikipedias and other, linked or searched sites as well.

tl:dr please don’t (unintentionally?) accuse me of (accidental?) copyright infringement, please :​)

editing to say that I don’t have an issue with the wikipedia blurb (generally, at least) but as I understand it, mfmeulenbelt’s issue is that the disambiguations aren’t translatable/search won’t show wikipedia excerpts. yes?

~Cat

2 Likes

Yeah, it depends on what “rewording” means in this case, and I haven’t looked at the new instruments. But I know, for older instrument entities, that they are often 1:1 copies from English Wikipedia; and I know that certain “translators” similarly take the Wikipedia text in their language and set it as translation, even though that means it often doesn’t correspond to the original English instrument description. (Note, not saying that @mfmeulenbelt is among those!)

Most of instrument names are language-dependent. Instruments are almost part of user interface. This prevented the use of untranslatable disambiguation field for instrument.

The description field is translatable and displayed in the editing interface, but it has many issues:

  1. It is not needed as a description anymore but to disambiguate, hence the other issues!
  2. Its name should reflect its real usage: disambiguation.
  3. Its current content is redundant with 1:1 Wikipedia articles displayed in instrument page.
  4. Its current content is unneeded to most of instruments which have very distinctive names.
  5. Its current content is not phrased as simple disambiguation comments, e.g., most of current English descriptions start with “The <instrument_name> is an instrument that” which is completely unnecessary/redundant in the editing interface.

(I did not list the raised copyright violations issue, since short descriptions are just necessarily using the same words, e.g., woodwind instruments are woodwind instruments, and so on.)

Here are a few non-exhaustive examples that need a disambiguation:

  • accordina/accordion (with spelling variants depending on the language);
  • cornamuse/cornemuse (in French only);
  • bass (catch-all for bass guitars and double bass);
  • tar/tar (with spelling variants depending on the language).

There is also the distinct case of instrument families which are usually named with a plural mark, like guitars and violins. They need to be disambiguated from their specific homonymous (singular) representative. I wonder if an additional “family” flag could help in this case. (?)

Rather than removing the translatable disambiguation (description field), we could:

  1. rename/relabel the description field to disambiguation (or any synonym like clarification if it hurts implementation constraints);
  2. drop the many now unneeded descriptions (that is, a vast majority);
  3. adjust the only needed descriptions to fit no more than their fresh disambiguation role.

It is a lengthy task, as lengthy as moving every worthy text to the disambiguation field, with the additional benefit of translation. I would happily contribute to this task for the French language.

4 Likes

Yes, all of this.

And especially this:[quote=“Yvanz, post:13, topic:105752”]
Rather than removing the translatable disambiguation (description field), we could:

  1. rename/relabel the description field to disambiguation (or any synonym like clarification if it hurts implementation constraints);
  2. drop the many now unneeded descriptions (that is, a vast majority);
  3. adjust the only needed descriptions to fit no more than their fresh disambiguation role.
    [/quote]
    It is going to be a lot of work to go through all these instruments (we currently have 786 instruments already!), but I think there are quite a few people who are willing to do some heavy lifting.

We would have to write English descriptions/disambiguations/whatever-we-call-it first, and then translate them through Transifex.

I think we can agree on the following:

Having a longer description duplicate WP is bad.

A short description (one sentence or less), to show while selecting an instrument, is useful. Fetching the first sentence of a WP article could work, but it’s often not descriptive enough, or redundant.

We could use the current description field for this short description (or whatever we call it). This has the benefit of translatability over using the disambiguation field.

1:1-cribbing from WP is bad, but especially with the proposed short descriptions there are only so many ways to express simple facts. I wouldn’t worry too much about it.

So, what is my proposition?

  • Use the description field as before, but take care to keep the descriptions short, and not simply copied from WP. Heeding the first, will often automatically disallow simple copying.

  • Only show the description on the instrument page if no WP article is available in the „right“ language. This code change is low priority – currently both the description and a suitable WP article are shown, which is pretty OK.

2 Likes

Depends, really. WP supposedly states facts, so the only copyrightable about its text is the specific wording or expression of these facts. So rewording may very easily make it a non-violation.

Unfortunately not that easy for people choosing to or having to use Tor. No open proxies - Meta
(Just a minor point…)

1 Like