I was merely pointing out that there have been many similar topics and feature request aimed fairly directly at classical navigation that have “apparently” fallen of deaf ears. Maybe search the forum for “Box Set” just as a “hot topic”. This AND OR debate is equally not new and the potential to upgrade the Focus tool has been mooted before.
I do apologize if my comment seemed flippant , that was not the intention . Many of the respondents above have been equally vocal of the classical shortfalls.
No offence meant and certainly no attempt to lessen the importance of your suggestion I for one would welcome the AND clause with open arms . To be able to go Composer then filter AND Concerto then filter AND Piano would answer my wildest dreams
I still don’t get it.
If Beethoven Piano Concerto No 3 works, that’s much easier, right?
And if it doesn’t work, wouldn’t it be better to fix that, than to provide tools that sharp people can use to build a solution?
“Beethoven Piano Concerto 3” does not work acceptably. “AND” is not a tool for sharp people to use, unless you think that people who can order a mac and fries are sharp.
My original request was to add a set of Genres (even Sub Genres if possible) buttons to the composer view such that you get the 452 Beethoven Compositions then click on Concerto to filter it down somewhat (and the same with the artist view).
I don’t know if my use case is different but I rarely use the Windows Roon app I tend to sit in my listening “Chair” with an iPad so any large amount of typing is quite painful on the soft keyboard I only us a normal keyboard when I am editing something on the core PC. With far too may PC induced ailments I am determined to minimize my keyboard time , hence Roon.
So the original concept was to set out the AND conditions as a series of “taps” to get from a large number of compositions (wort case is 1200 for Bach !) to a manageable set for final selection of the composition and hence the specific recording artist.
Whilst I agree typing Piano Concerto No.3 in the “Funnel Filter” is possible its not the behavior I want to use. My old system (JRiver & JRemote) gave me literally the series of taps I want with absolutely no typing, why can’t Roon , an overall more slick experience, give us a similar typing free experience.
I suspect may of us have spent far too much time at a keyboard in the past so my vote is for a typing free zone
2 Likes
James_I
(The truth is out there but not necessarily here)
47
Foobar’s Columns UI view does this extremely well. Each column is a filter and you get exactly what you want remaining at the right-most column:
You refuse to pay attention to a word I’m saying.
I don’t think people who can order Mac and fries are sharp.
This is not about what I think.
I have never said anything about what I think is easy to use.
I am referencing 50 years of industry history.
I can’t think of a single product with significant impact in the consumer market that has used Boolean logic.
Lots of products have had huge impact in the consumer market, and all of them use intelligent natural language, unstructured search.
Of course, all products aimed at techies use some variant of Boolean logic.
So this is about what market segment Roon wants to go after.
That’s a business decision.
It is not about whether I would find that feature useful (and I have never said whether I would).
And if I may be so bold, it is not about whether you would like it either.
It is disingenuous to claim that a feature is obviously good for everybody, without acknowledging that it involves a choice.
I first saw this in Smalltalk, which in 1980 did an object-orientation similar to Simula 67, which I first used in 1973.
Ah, the 1970s and 1980s, those were the days.
James_I
(The truth is out there but not necessarily here)
51
Anders, I enjoy your posts. But what’s your point here?
That Roon has come up with a more modern way to not make their Tags function useful and thereby prevent curator-users from doing things we have been able to do with other media applications for years, even decades?
You’re having two arguments here: Boolean logic may not be the best choice to use with Roon’s search function (although many may find that useful) but in terms of Focus and Tags, it is a natural way to identify the intersection in the Venn diagram between two sets of Tagged music. If you can think of a better way to quickly accomplish that, I’m listening, but otherwise, are you just here to take pot shots?
EDIT: One of my clients is being truly annoying and I vented that out in this post. Anders is entitled to his opinion. I do think the comment needs a conclusion or a more directed point, though.
I think the problem is not the search algorithm but the tagging. You can’t do a full text search of musical content like you can with documents and web pages. The ontological approach of Classical —> Concerto —> Violin should yield the same result as the intersection of all three and not the union I’d the data in the meta tags is sufficient. Even Google can’t add data that’s not there. Having a tool to edit the tags yourself and share them with other people like the Gracenote database if the people doing the editing and sharing know enough about music to do so accurately. Bad data is just as bad as too little data. If there’s a site that has an accurate ontology of musical genres that would be helpful, but it has to have specific examples to be of any use here. Parsing every site related to music and using machine intelligence to create and accurately apply tags is probably not an NP complete problem because language changes all the time, new genres emerge, and albums come and go. It might help if Roon incorporated a thesaurus into Focus to help disambiguate terms, but that assumes tags come from an enumerated set, and tags are almost always free text. Roon really can’t do anything about data. Plus, the engineers at Roon are very talented but I doubt they have any search expertise or any practical knowledge of data science. That’s usually used for marketing by companies like Pandora, Spotify, and TIDAL that sell content, but Roon is a software-only company. YMMV.
No potshots.
It’s just my long-standing concern about this long-standing request, which I consider not always well considered.
The discussion has gradually converged on a more well-considered approach, use of such logic with tags only, specifically in the interest of curators (no, I’m not going to claim any credit for that evolution ). But that is a shift: even this thread, it’s title and first post, were about generalized Boolean.
With this approach, we explicitly say it is aimed at curators and their custom tags, which they presumably know. And a community for sharing such data could evolve. That’s a concrete plan that can be evaluated: is it plausible, is it big enough, does Roon want to go there? Cool. Good luck.
There have been several specific market segments that Roon has chosen to go after—the HQPlayer- and Computer Audiophile-style filter tinkerers—or not go after—the Spotify “don’t bother your little head with choices, we’ll take care of the playlists for you”. (He ducks!)
The more concrete the idea, in technical and use case and audience terms, the better the chances of adoption and success.
James_I
(The truth is out there but not necessarily here)
54
OK fair enough - your comment is really more about the process of prioritizing software development. See my EDIT in the prior post.
Tags are simply too one-dimensional without this capability.
Not wishing to labour the point, this is how another product allows this logic .
In this case a custom designed view based on Panes , One click per pane takes you down to a single recording of the Composition with NO TYPING. Albeit with an enormous amount of manual effort to populate the tags in the first place.
The same view can be used in a iPad app to achieve the same effect but without Roon’s rich appearance.
Surely this should be feasible with Roon’s rich metadata sources
The main difference obviously is there are no drop down lists and panes. But arguably there are advantages to the roon approach in that there is less preparatory tagging overhead involved as in the “panes” example (I am playing Devil’s advocate here). On the other hand with a larger collection you will almost certainly have to “merge” a lot of compositions, otherwise roon doesn’t really work and there will be many orphan compositions that will not show up in searches. The jury is out for me whether merging in practical use is easier or harder than tagging.
Roon method:
Go to compositions.
Type in “bee” in the composer column filter. Confirm if you see enough Beethovens.
Type in “piano co . . .” in the composition column filter. Confirm if you see enough Piano Concertos.
Scroll down to Piano Concerto No. 3 in the Compositions column and confirm.
Go to the “Piano Concerto No. 3 in C minor, Op. 37” composition screen. On my roon I see 4 versions of this work in my local library and 242 in Qobuz (a distinct advantage of roon over JRiver).
I don’t have the Argerich/Abbado version and the Qobuz list is way too long to scroll through.
Type “arg . . .” in the filter at the top right and you will see the Qobuz list converge until you can pick out the Argerich/Abbado version.
If you wish you can “add” the streaming version to your library and edit local library tags to taste.
It is arguable which is the better way of doing things. It seems to me there are pros and cons. Maybe the roon ergonomics could be better. A big problem with roon’s heavy graphics and lack of dropdowns in this scenario is that there is more typing involved and when you are “searching” rather than “browsing” this all gets in the way. At the end of the day I think the roon approach is more geared to browsing than searching and that might be what is at the root of the frustrations (which I also share).
Your method does not work unless you have the work you are searching for in your library, so it is no use for compositions or composers you don’t have in your library. It might have been useful when Roon was closed and restricted to your own rips, but now Roon is open to the vast resources of Qobuz and Tidal the search needs to be revisited and improved.
I know and I use exactly that method myself but it’s all typing , the panes view look is all mouse clicks or taps
I’ll hang in there and wish , I still use JRiver actively for this reason if it’s simply to select and play an album/ composition
The Tidal integration is really all that stays my hand, it’s so frustrating even more so when there is no apparent comment from Roon. We can be excused for thinking they are ignoring the issues
In practical terms what would you suggest to improve the (largely) hidden boolean features of the existing interface? It must be obvious by now that roon will never provide a boolean driven list/pane interface. This would appear to be beyond a logical or practical issue for roon. It looks like yet another point of principle like DLNA.
). But that is a shift: even this thread, it’s title and first post, were about generalized Boolean.