[eschwartz]

Originally Posted by Faterson

I disagree – it's extremely hands-on and practical. When you see multiple listings for a single author in the comprehensive authors listing in your Library, this is a very practical alert for you to fix your meta-tags. What more could anyone wish for? Do you really believe that a user would prefer to see authors sorted according to their first names instead, even though he/she checked the "sort by last names" preference in Marvin settings?

The "best solution" I was referring to is, of course, implementing a vastly complicated algorithm to properly derive the right sort data from the author field, which, if it existed, would mean no one would have ever created a sort field to begin with.

Every other solution requires you to sacrifice something. And while you might prefer the awkwardness of seeing duplicated authors (because correctness is the prime virtue) kguil has to think of what most users would prefer to see.

And while I respect and sympathize with your preference (being a compulsive metadata correcter), I absolutely refuse to entertain the notion that you automatically know what everyone would prefer. That is a rather ambitious claim!
In fact, the average person has proven to be below average. () And Apple aggressively targets those people -- so maybe it makes sense for iOS app makers to do likewise. Not saying that is the case here, though it might be.

Nope, that wouldn't work at all. Many authors have several last names, and many books have several authors.

And yet, that seems to be exactly what calibre does... several last names or not.

As for several authors, the author list will already be splitting those for each author listed separately in the OPF.

Why would you wish to re-invent the wheel, eschwartz? Everything Marvin 3 needs to produce reasonable authors sorting already is available in Marvin 3: namely, the (unfortunately labelled) "author sort" meta-tag field. That is all that is needed. The unacceptable flaw in Marvin 3 is that despite Marvin 3 having such a meta-tag field, it chooses to ignore the data entered in that field in the comprehensive authors listing. And that's a big no-no, as I believe even you admitted.

I have admitted it is not ideal, but I have also stated that I can understand if someone says it isn't practical to implement.
Which seems to be kguil's belief -- so why don't we wait for him to get around to writing that blog post? Unless you believe continuing to assert your counter-claim will change his mind?

Well, all I know is that the fix I'm proposing does not come with any performance penalties whatsoever – because it's a trivial fix, fully logical and reasonable, and because I simply wish for Marvin 3 to work like Marvin 2 and Marvin 1 in terms of authors sorting: allowing us to sort authors either by their last names, or by their first names – whichever option the Marvin user prefers. And this is perfectly possible even with the current menu options in Marvin 3 – if only Kris finally agreed to put the meta-tag fields, primarily that (unfortunately labeled) "author sort" field, to their proper use!

You don't know it is that simple, you assume it.
I don't know either way, and I am not going to assume.

It is not always that simple.