To be honest, lately I’ve been thinking about this from the viewpoint of, “just give me something like NVK Search for everything, and then I can just be done with browsing modules for the most part” 
4 Likes
That’s pretty much how I expect the search to work as well. However that ONLY covers the scenario where I know the module name, or at least brand name. So it can’t be the only implementation of the browser.
I imagine this would match based on categories as well, so if you write something like “c filter” it would match every filter, and so on. But yeah I agree, naturally this wouldn’t be the only way to browse.
Maybe its suitable for the text based browser
That’s my dream, basically. Personally, I wouldn’t usually need to see all the different module graphics in the browser like it is at present. Something like a text-based mode with a powerful search function like above, and having a dedicated part of the window for showing a single module graphically when merely mousing over some particular name… would probably be optimal for me in at least 95% of cases. Maybe more. Lol.
1 Like
Have you tried Prefabs by ldlework pre-release? It’s almost exactly what you describe.
It’s not in the library but @ldlework did a great job implementing a fast text based search across modules and selections with only graphics if you hover over a search result. It even has a replace browser option 
4 Likes
Whoa, haven’t seen this! Thanks 
couple of Q I have.
-
What would a text mode browser look like exactly? What info does it show per module? Can you controls what if shows (like in windows explorer)?
-
What exactly is “powerful search”. Can it be defined in such a way that it could be implemented?
Heh, not trying to avoid the questions here, but the week started with a load of things to do in work projects 
so I’m just being brief/vague: the above video of NVK Search gives a pretty good idea of the vibe I have in mind (I’m also an avid user of Everything search by Voidtools, and FSearch on Linux; also the feel of fzf and integrating it with ripgrep, ripgrep-all , and so on, are other good examples of the raw feel/vibe in a completely different context).
In the case of VCV, a search like this should be able to target at least the module name, brand, version, author, description, tags/category, probably also update time (at least for the purposes of sorting exact-word-matching results), and so forth. A “powerful search” can fuzzy target these information categories in the manner displayed in said video, and the results would be presented in a list with - for example - user selectable columns for information associated with modules, the main information for a default entry in the list being the module name and brand. Also, a specific info area, in a part of that search/browser window, which graphically shows the module that is mouseovered in the textual list at any given time, along with infotext and the like, would be very nice.
Just for the record, this isn’t a feature request, just thinking out loud what kind of browser/search (edit: browser/search mode ;), not saying “get rid of the all-graphical one”) would be a bullseye for me personally :-3
When nothing is entered in the search field, display all modules (either literally ALL modules, or when so selected, all modules in favorites / selected subcategory / etc.) in a list sorted by chosen criteria. Narrow down this list accordingly when something is entered in the search field.
2 Likes
all good stuff. I think there use to be a post from Andrew at VCV - something like “request for better search algorithm”. so I don’t think the issue is convincing VCV that a better search would be nice. It’s more that they don’t know what a better search means, or how to make it.
fwiw, I think the idea of a text only display is great. Of course as I mentioned I think every single product that lets your search for stuff has that option (in particular windows file explorer). That thing has been evolving for the last 30 years, at least. So let’s see what one for VCV would look like!
Hey, by the way, do you happen to know what happened to this project? It seems it was going along really well, and the author said they are going to refine the implementation etc, and “try to make a pass for stability and then release 2.0.6”, but the current build is 2.0.5 and from over a year ago, it has just stopped cold with no messages (from anyone) afterwards?
I hope he stopped being idle and got some real work done 
But do be honest, I don’t know. I believe he was quite active on this forum and on the VCV discord so perhaps someone else knows.
The code itself compiles and runs on the latest rack api on linux as far as I can tell. I saw there’s problems on arm though.
I tried to modify it some time ago so I could type
t:tag1&tag2 n:name1|name2 b:???nd|br*
And a good coder like Dustin could probably type that up in minutes But it took me too long and gave up.
You could ask him this question on his github I suppose.
The current Rack module browser has a fuzzy search algorithm that it uses, which was contributed and can be reused.
My biggest gripes with the browser is usability. It’s kind of in line with lots of other software I see these days, which seems to be only designed with trackpads in mind, and always assumes a good scrollwheel. My strong conviction here is that both mouse and keyboard should always be first-class citizens and in that respect the browser is kind of user hostile.
So, when you have a long list of names/items it must have a scrollbar and it should have/be a combo-box, where typing narrows the result on the fly. This used to be “of course” knowledge 20 years ago.
I would personally also like a fast, text-based browser, with some of the features mentioned in this topic, where “show the image” would be one of the keyboard shortcuts. Fast narrowing down of results using fast addition of search vectors (brand, plugin, tag, …) also with the keyboard should be king, and would be a great experience, at least for power users.
“Power users” is an amazing term lol. I also think theres a lot to be said for a set of features that allows users to customize the types of experiences they are looking for.
This can be as simple as text based vs graphic browser all the way to as complicated as choosing features… (sorted clickable tags, etc)
I am a user highly invested in patching speed for live patching and performing, as well as recording on other peoples work… kindof niche, but a “modular” browser solution would allow users like me to dial in the type of experience im looking for. Building for that in mind might not mean significantly more work neccesarily…
2 Likes
Well said, a few customizable tags or folders and we can all be happy happy. Can’t be that hard.
Usability! Yes. I don’t know if you’ve ever participated in a “usability study”, but it can be a humbling thing. Although these studies can take many forms, the basic one is hand someone your software, ask them to perform a simple task, see how horribly painful it is for them to accomplish the task!
More on topic, yes, most good software will have a way of using he software with keyboard as well as mouse. It’s probably becoming less common now - for sure us “olds” have used or created software in the time before a mouse / pointing device was common
Still the ideal is to make your user interface work well with only a keyboard and also with only a mouse. And “work well” tends to mean that all features are usable, and that most workflows can be accomplished with as few gestures as possible. Especially the most common functions “should” be accessible with the absolute minimum number of gestures (or “click”, as the youngsters call it).
Also, as we are seeing on this thread, different people may have different workflows that they really care about. It’s “easy” to make software work well for one person, it’s very difficult to make it work well for “everyone” or “most users”.
5 Likes
It sure can and it’s good for the soul.
I’m one of those appearently rare people in the IT biz that can both build software and immediately turn around and be a user, empathize with the user and see exactly what it’s like from their standpoint. I care deeply about end users and find no point in it otherwise. It took me a long while to realize that for most developers that’s very difficult, don’t know why. I also appearently have a high level of allergy to bad usability and can usually find at least a handful of usability problems in any piece of software you sit me down in front of, within the first 10 minutes. I guess that makes me odd and annoying
My belief is, unfortunately, that software in general have been in steady decline with regards to usability for many years and it continues. By my standards and rough memory, usability standards reached a peak somewhere around 1995-2000 and have been declining since then.
haha - same for me. But I did have to stop caring too much at work. In VCV it’s ok if ppl hate me, but at work I have gone too far being too negative and critical. It would be bad for me to get along poorly with my co-workers and managers, and even worse to get laid off or fired! So I happily make whatever ppl ask for at work, and use VCV and other side-projects as a a place where I can make stuff just like I want.
Oh, wait, I have been suspended from VCV before! Well, I try to avoid that now and behave at least tolerably here.
3 Likes
Yeah, it’s tricky business, I know exactly what you mean.
2 Likes
BTW, for anyone interested in usability, the old CUA standard and the Nielsen guidelines are very good.
2 Likes