Besides that it would be wonderful to have access to the main alias for scripting, so e.g. by tagging a release from AFX (Alias of Aphex Twin), I could create a symlink to an Aphex Twin directory like this:

Artist/Aphex Twin/Aphex_Twin-2001-Drukqs
Artist/Aphex Twin/Aphex_Twin-2014-Syro
Artist/Aphex Twin/_Aliases_and_Variations/AFX/AFX-1993-Analogue_Bubblebath_Vol._3

I just noticed that on Discogs there is a difference between Alias and Variation, maybe because of that, the files are tagged wrong.

The alias name appears in the data stream if the data was entered correctly. Storing the name when available is no problem.

The issue is that there can be an alias' for every artist name. ie/ multiple alias' for a single release.

Perhaps storing the alias' as: Discogs_Alias_for_this artist=alias name makes sense. This would be a UDTI which could then be moved to any other field. I tend to like this approach because given the 'artist name' you could construct the UDTI name and check for existence.

I'll look into this as part of the previous 'alias post' and will get back to you at the same time.

In the case of:
http://www.discogs.com/Cosmin-TRG-Repetitiv/release/5823744
The situation is:
name=TRG alias=Cosmin TRG
The web site displays: Cosmin TRG* ‎– Repetitiv

Here's the raw data:

----
artists: Array[1]
artists: [0] Dictionary
artists: [0] tracks:
artists: [0] id: 567073
artists: [0] join:
artists: [0] role:
artists: [0] anv: Cosmin TRG
artists: [0] resource_url: http://api.discogs.com/artists/567073
artists: [0] name: TRG
----

In the case of:
http://www.discogs.com/Doms-Deykers-Fonts-For-The-People/release/6021864
The situation is:
name= Steffie Doms alias=Doms
name= Martijn Deykers alias=Deykers
The web site displays: Doms* & Deykers* ‎– Fonts For The People

Here's the raw data:

----
artists: Array[2]
artists: [0] Dictionary
artists: [0] tracks:
artists: [0] id: 2119321
artists: [0] join: &
artists: [0] role:
artists: [0] anv: Doms
artists: [0] resource_url: http://api.discogs.com/artists/2119321
artists: [0] name: Steffie Doms
artists: [1] Dictionary
artists: [1] tracks:
artists: [1] id: 518875
artists: [1] join:
artists: [1] role:
artists: [1] anv: Deykers
artists: [1] resource_url: http://api.discogs.com/artists/518875
artists: [1] name: Martijn Deykers
----

My conclusion is that the Discogs web site displays the alias names as opposed to the 'names'. I could put in an option to use the alias names when present, however I cannot guarantee that Discogs does not sometimes present an alias when the actual 'name' is correct. Nor can I guarantee that there is not more than one alias given for a single artist.

BTW: in DIscogs terminology both the above examples use ANVs and not alias'. 'Also Known as Values'. To make matters worse there are also alias values.

I can save the alias names (ANV). We just have to decide on a format which would work. They can be saved one User Defined Text Item (UDTI) per alias or combined so that they can be iterated or if you really want, both ways.

eg.
Discogs_Alias_for_Steffie Doms=Doms
Discogs_Alias_for_ Martijn Deykers=Deykers
or
Discogs_Artist_Name_for_Doms=Steffie Doms
Discogs_Artist_Name_for_Deykers=Martijn Deykers

or:
Discogs_Alias=Steffie Doms=Doms;;;Martijn Deykers=Deykers
or
Discogs_Artists=Doms=Steffie Doms;;;Deykers=Martijn Deykers

Opinion?

As for arrays. Introducing data types in Yate would be very tough as there are none now except for 'string'. However you can use the 'Repeat With' function to iterate over delimited data. There are two choices:

1) You can uncheck the Prefs-Discogs-Combine multiple values into a single value option. In this case the artists and all array based fields will be delimited by the multi-value separator (;;; or \m).

2) If you prefer to leave the 'combine' option checked you can 'repeat with' the appropriate delimiter. The delimiter is almost always '&' as you can see in the above raw data under the 'join' parameter. If a 'join' is not supplied, Yate assumes '&'.

So the feedback I'd like is:

- add an option to default to alias names when present?
- what format preferable for alias information?
- save the artist information when alias/artist is flipped?

Sorry for the long rambling answer. I'd like to end up with something that makes sense 🙂

Wow, thanks for the quick implementation!

As for arrays. Introducing data types in Yate would be very tough as there are none now except for 'string'. However you can use the 'Repeat With' function to iterate over delimited data. There are two choices:

I did not mean new data-types, just arrays as separated list or multi-value-field.

Just checked with an Aphex Twin alias (http://www.discogs.com/Brad-Strider-Bradleys-Beat/release/3579). «DISCOGS_ANV_LIST» absolutely makes sense. Not sure, if «DISCOGS_ARTIST_LIST» is needed.

Sadly (due to the Discogs API), the ANV in this release is not «Aphex Twin», but «Bradley Strider», so linking to «Aphex Twin» as ANV would require a lot of additional work, probably going through all ANVs and checking the numbers of releases to find out which ANV is the most famous one. Seems like the ANVs also vary, depending on which artist side you look at.

I opened another topic (http://2manyrobots.com/forum/?mingleforumaction=viewtopic&t=220) about separators and multiple-value-fields. I think it would be good to have separate editor fields, a multiline field (comment), a multi-value-field (multiple artists) and a raw field (which shows escape sequences like «/0»).

I don’t really understand how Yate handles multi-value fields and separator chars. «;;;» and «/n» are non-standard Yate only separators, right? ID3 uses «\0» or «\» (http://en.wikipedia.org/wiki/ID3#Multiple_values_in_v2.4_tags). Not sure about Flac. I’m just a bit concerned about standards. There is a website (http://sonicsquirrel.net) which cannot use Yate-tagged files (not sure on which side the problem is).