I don't agree lowering (or raising) the Masternode collateral limits just to please the "poor guys" or the "rich guys" in the community, because that would make no sense at all.
This must be a technical decision: what is best for DASH performance.
I agree with the spirit of your point but I am not sure what you are implying by the details, so I would like to expand on something.
I agree entirely that making changes to Dash just to please one interest group or another would make no sense, it would be making changes without logic, and would set Dash adrift being pulled one way or another by each group. What I think needs clarifying is the scope of "technical decision" and "DASH performance". It is easy to read these terms and think of only decisions that directly relate to code, and only performance that can be measured by hard metrics. I would like to expand first the scope of "technical" to include everything that can be managed scientifically, even if it is hard to measure directly. So this would include management, marketing, purchasing (eg paying for an ATM integration through the budget system); second the scope of "performance" to everything relevant to using Dash, so how quickly a shop customer can buy their coffee with Evolution, as well as how many InstandSend transactions per second a masternode can validate. (I only just wrote a
comment on Reddit where I use the same ideas.) It is much harder to measure the effect of an social media promotion on Dash usage than it is to measure the effect of a code change on block verification time, but here are still technical decisions (eg "don't be rude") that affect performance (eg "percentage of people who read tweets who then visit dash.org"), even though they may be harder to define and measure.
You may have thought about some or all of these already – I just try to expand the scope of "technical", "performance" and so on to cover as much of a system as possible, otherwise there's a risk of focussing too much on one part and forgetting about another. Everything works together to create Dash, and changing one thing affects everything else indirectly, so it helps to keep a wide field of view.
But, a possible solution would be that of the spreadcoin guys: no fixed limit for the collateral, but else, a fixed optimal number of Masternodes....
...Say the developers calculate that the network needs 5k Masternodes in order to operate perfectly in the current context. So the network will only recognise 5 thousand Masternodes, and there will be a set of minimal requirements for a Masternode to be accepted, like reliance, speed, responsiveness, ping and whatever is important according to the technical point of view.
This sort of line of thinking is promising I think. The 1000 DASH is just an arbitrary figures should that has proved a good starting point, but the number of masternodes would be best determined by fitting into some broader goal for Dash. How, exactly, I don't know – but I think it's worthy of further thought and research.