Since you say you're a programmer, I see where criterion #1 comes from. But telescopes are not computers, you can't upgrade the CPU today, the RAM tomorrow, and so on. A scope is defined largely by its aperture (the diameter of the objective lens or mirror). That puts a major cap on pretty much everything else, performance-wise. Aperture is like an old boarding school taskmaster who says "you're allowed up to here, no more", and anything else you may do can only place you lower than that ideal level of performance.
Scopes optimized for visual and scopes optimized for photo are different animals. They are interchangeable to some extent, but after some point their respective traits start acting up. Usually, people start with a small price-efficient visual scope (like a small dobsonian), then migrate to astrophoto after some learning is done. But if you're intent on doing AP directly, fine.
The instrument that's typically used for AP is some kind of catadioptric, like a Schmidt-Cassegrain (SCT), on a tracking mount. It doesn't have to be an SCT, it could be a refractor, a Ritchey-Chretien, it could be a newtonian, or what have you. But an SCT is typically short, stubby, and rather not unwieldy for its aperture, which are good attributes if you put it on a tracking mount.
The mount doesn't necessarily have to be a go-to mount. I would argue that go-to is a waste of money if you're smart enough that you can use a star map (either paper, or software). But it absolutely needs to track the motion of the sky on one axis, because you're going to take a lot of long-exposure photos.
As you can see, AP scopes and mounts are expensive. You could get away with a mediocre scope, and still take good pictures, but a bad mount is a deal breaker.
A lot of beginners are like "okay, that stuff is way too expensive" and simply purchase a small Cassegrain on a cheap go-to. It's definitely less expensive, but the performance changes accordingly. Some examples:
Finally, if you decide to take a while and school yourself in purely visual astronomy before purchasing a killer AP rig, you could start with a bang-for-the-buck visual scope such as a classic or Intelliscope dobsonian - as much aperture per coin spent as possible:
(I'm in the US, but the general principles should apply; maybe less so the particular examples above.)
EDIT: Here's a bare bones rig for AP:
If you have a digital camera already, it will cost you almost nothing, and you could assemble it in one week-end.