Adroit Custom News

[ColinP]

As you are probably aware Adroit Custom is well behind schedule. I came across serious problems just prior to putting it into beta testing. Should you be interested, you can read about some of the technical details here... viewtopic.php?t=2957

I've pretty much finished a new user interface system that fixes the problems but rather than doing a quick mashup with my old code I'm going to do a third rewrite as the existing code is so complicated that it needs another thorough clean up. When a project approaches about 10,000 lines of code then productivity plummets as one has to constantly relearn how the existing code works before modifying it. So it's often more efficient to rewrite everything in order to remove all the ad hoc, half-baked solutions that have accumulated. A good rewrite usually halves the amount of code and results in everything being clean, efficient and easy to modify.

Anyway, as part of this rewrite and now having almost full control over the user interface I'm left pondering about which direction to take on the skinning aspects of Adroit Custom as redesigning Custom Look is a realistic option given the amount of work I'm doing.

There are two contrasting approaches. For instance in the case of knobs, one approach is to have a small set of predefined knobs to choose from. Alternatively users could import their own knobs. The first approach is easy to use but inflexible while the second is difficult to use but very flexible. Of course one can have a mixed approach but that makes things even more complicated.

Previously Ive been thinking that the predefined approach is best but I'd be very interested to hear any feedback about how many people would like to do their own graphics design (although supplemented with a small library of graphics files to get going and to effectively offer the predefined route but with a bit more faff than a purely predefined setup).

In practical terms a knob consist of two files, one containing the fixed part of a knob and the other the rotating part of a knob. So as a minimum you'd need to select the two files from a folder. While to create your own knob you'd need to create the files yourself.

I'm wondering how many people would have the inclination and skills to do this kind of work. I'm using Inkscape to do this stuff and I reckon it takes roughly 20 hours effort to become moderately skilled with such software and maybe three or four times that if you are not already familar with graphics editing.

I realise that few people actually respond on this forum but your thoughts on this would be very welcome.

[UrbanCyborg]

I'm one of those people who would likely design my own knobs. I've already dipped into modifying existing knobs (see the recent update to my Poly Voltage for an example). That said, I'm also one of those people who has a hard time sequestering the time necessary to do so.

Reid

[utdgrant]

I would definitely stick with a small (initial ? *) library of pre-defined controls. Yes, it goes against the 'modular ethos' of being able to configure and bend EVERYTHING to your wishes, but purely from a user's perspective, I think it's one step too far down the list of things I care about.

I liken Adriot Custom somewhat to Combinator within Reason. Combinator (and particularly Combinator 2) is great for hiding a rat's nest of modules and cabling below a concise and slick-looking front panel. However, I've never felt the need to design my own custom controls. Similarly, I'm not (extra) impressed by Combinators that other people create when they've gone to town on the aesthetics. Just give me the functionality (and explain it, please)!

Many people are too lazy to even document what the available controls do! So I'd be surprised if they'd want that level of detail for the look of a custom knob. Certainly not in comparison to a 'standard' one which has been lovingly designed by the manufacturer to look great and function well stright out the box.

This is just my personal preference (as a user, not with my developer hat on), and so should absolutely be treated as just another single data point.

(*) Maybe add refill packs down the line whenever the fancy takes you?

[ColinP]

Thanks for the input guys.

I think the time element is crucial here. While a graphics designer could easily knock out a really nice knob in under an hour it might take a non-professional all day to achieve something that's nothing like as polished.

And yes, you're right Grant - I've also seen some truly awful DIY graphics (not particularly in Combinator, but in general).

Perhaps preset components are the best option and giving users full control over the background and font choice is sufficient.

But preset libraries have their problems too. For instance the most practical, although rather boring, stock knobs in VMD are the "plastic" ones but the color choices are quite limited and asymmetrical in color space terms.

PlasticKnobs.png (64.64 KiB) Viewed 5952 times

To make this flexible one would have to have a very large number of presets.

A third-way I've thought about is programmatic. So one has a very limited number of styles but the basic color of the knob etc could be adjusted using Hue, Saturation and Brightness knobs. This could be implemented by image processing filters so there'd still be underlying flexibility in the style choice (even if this choice is controlled by moi rather than end users).

It would be good to hear a wider range of opinions, perhaps someone from the other end of the spectrum? Calling all you budding artistic types...

[drohnee]

I would vote for either allowing users to create their own files as long as a few template or examples are provided so it is obvious what needs to be done, or else have a few basic shapes and allow users to select colors with HSV or RGB color picker. Even though I am certainly no visual artist, I do still find that I have preferences about the aesthetics of things like this, and if I am putting together a custom UI panel, I will probably want to at least control the colors of the elements.

Even though VM is in the skeuomorphic design camp, I would imagine some folks would like to be able to create UI panels that have a more flat / minimalist aesthetic, and being able to do this with Adroit Custom would be a plus in my book. On the other hand, I think the argument that keeping things as simple as possible for users is a valid one, but I tend to think giving more choice is a good thing.

[ColinP]

Cheers drohnee. It's really good to see a mention of skeuomorphism. My tastes oscillate on this. Having just spent about eight frustrating hours trying to add efficent yet believable shadows and specular hightlights to my joystick code I'm currently a big fan of flat minimalism.

I quite like a mixed style where UI elements float mysteriously above the surface yet still cast shadows and have lighting effects. So a hint of physicality while still being abstract and uncluttered.

But everyone has different tastes and it's nice to have a choice. Aesthetcs might seem like a minor concern but we'd all be wearing gray overalls and live in beige cubes lit by flickering blue/white florescent tubes if it didn't matter. On the other hand few of us make our own clothes or design furnishings.

I can't get used to calling HSB HSV. Brightness means something, value doesn't but then I'm an old git who's being using HSB for almost as long as RGB.

Anyone interested in color space might find this wiki page enlightening... https://en.wikipedia.org/wiki/HSL_and_HSV