Ok, I’d love to see a modern-day Arp 2500.
Make it scaled down slightly in size, use electronic switching and bussing instead of the slide switches, realize all the sound circuitry with mostly discreet parts (transistors instead of opamps), but then a CPU that does a background “watch dog” function to keep everything autotuned, along with high accuracy ADCs and DACs to make it all programmable with recall. MIDI and external analog I/O, of course.
The pots could have over 16 million possible settings, for example.
No stepping or zippering.
You’d get a synth with nice tonal characteristics, pots and controllers with such high resolution you wouldn’t know they’re not analog, the beauty of the 2500’s multing system and general layout and none of the drift, unrepeatability or crosstalk.
Everything programmable too. Even program sequencing.
The slide switches could be color-coded square lit pushbuttons instead.
Have large matrices of buttons easy to push and visually identify.
That could make it really neat to look at and easier than looking down the rows of the 2500, one of it’s larger criticisms.
(If you’ve ever programmed a 2500, you know what I’m talking about.) 
Both original style 2500 modules could be duped and newer things like quantizers offered.
The original graphics could be included with arrows pointing to patch busses, etc.
The buttons (instead of slide switches) could be exclusive (only one lit at a time: original style) or be from or to multiple sources or destinations.
(press more than one at a time.)
Visual matrix routing ala Arp 2500, but brought into the modern day equivalent.
The downside would be the cost.
Something like this could only be realized via surface mount technology.
That doesn’t bug me and can actually increases accuracy when you’re talking about matched components, but it would be more difficult to service or mod.
The main enclosure would cost the most too.
It would be filled with high quality buttons, circuitry, connectors, power supply and the case itself.
So…a slightly smaller Arp 2500, lit like a Christmas tree, all programmable and great discrete tone?
Neat idea or silly dream?
Perhaps an open-source release that provides support for other companies to develop modules for it too?
A modular that could expand with time?
It could lead to other people making modules for it, so it wouldn’t become obsolete.
How many people would buy such a synth and at what cost?
What would it actually cost to build in a small number like 400?
Anyway, that’s my idea for something possible given money, talent and today’s technology.
No idea if it could ever be profitable, but I think it would be a really nice synth if done like this.
I have no doubt someone could make a virtual equivalent, but it just wouldn’t sound or feel the same.
Any comments? 
) opinion, considering the economy as it is, the way prices are going up, disposable income rapidly disappearing, and that people like us tend to be more of a niche market, I think it would be very expensive to produce, and likely have a limited number of buyers in the end. Not only that, but if it is billed as being an ARP 2500 clone, “the experts” are likely to b*itch and moan about every little thing on it that isn’t just like the 2500! They’d pick it apart to the point where some people might put off buying one.
) or by LED tracks which show various audio and CV routing. Imagine an LFO sending its waveform to a filter etc.. I guess this would kind-of be the modern (digital) equivalent to patch switches/matrices but in a more visual, intuitive way.
