Hello!
Apologies if this is in the wrong section, and I apparently have a few questions.

So I've been used to running one project on a variety of machines, simply changing the specific projects e.g. mapping cancer markers, that I wanted to use per machine on WCG.

Now I find myself floundering a little. How do I manage how many resources (100, 200, 40, 60) of each project? What if a higher resource project (say, 200) has no work yet the 20 does? Will boinc still grab some tasks? Or am I overcomplicating this? Should I just keep each project at the default and let Boinc do its thing with them? Wow, that was a lot of questions at once, eh.

I've got a few machines here that have 4 gb of ram. I think Climate prediction has something like a minimum of 1.5 gb per core. Is there a way I can find all of the shortforms for each specific project so I can somehow use the xml file to limit the number of WUs worked on at once if I need to?

In a similar vein, without going onto every single forum of each project I'd like to contribute to, is there a way I would be able to see each projects (or various types of WUs in each project) average runtime and what the deadline is? This way I'm not adding too many projects and not getting work done in time.

I think that's everything for now. I added a few projects to a Linux machine (Citizen science grid, Denis) that both got no work added. Similarly, RNA doesn't appear to have any work according to Boinc's event log. At least if a pile of work suddenly appears this machine's primary purpose is crunching so I'm confident it will finish in time.

Any hints or suggestion are beyond appreciated. I feel like I've gotten a little over my head. Or maybe I'm just overthinking this entire thing. Even going so far as to wonder if there are any specific projects worth running in linux vs. Windows and the other way around.
That being said, I'm going to thoroughly enjoy branching out and expanding my contributions to other projects and will enjoy the ride.

With more than one project being allowed to download work, the resource share ends up being a ratio. If one is at 100 and another is 20 then eventually one project will run 5x as more than the other. It's a long term thing though so once a 2nd project is enabled then BOINC will try to catch up its run time to meet that ratio. If you want to run each project evenly then set them all to the same #. It could be 1, 10, or 1000 if you want.

I keep a project or two at 0% as a backup in case the primary project goes down or runs out of work. Then it will only download enough work to keep the processors busy.

The scale for resource share is 0 to 1000 for each project. The total does not have to add up to 100, or 1000, or indeed anything else (but it can get very confusing if you set all projects to zero!!!!)
Zero has a special meaning - If there are no tasks available from any of my non-zero projects then get a task from this project.
Any other value is a weighting for how much effort should be expended on each project over a period of time (weeks, not minutes)

So, if you are running five projects:
A - resource share =1
B - resource share =10
C - resource share =100
D - resource share =1000
E - resource share = 0

Over time the work done for each project will be in proportion to the resource shares, expect for E. E will only do any work when there is nothing available for any of the other projects. You are correct in your assumption that if there is only work from B, and none from C or D then work will be done on B, until work becomes available from C or D. B will then "slumber for a while, until it has fallen below its target.

Any hints or suggestion are beyond appreciated.

One problem is that the BOINC scheduler takes forever (almost a month) to adjust the number of work units to the resource shares you set.
A way of speeding it up is to modify your cc_config.xml file (in the BOINC Data directory) to include:

<cc_config>
  <options>	
  	<rec_half_life_days>1.000000</rec_half_life_days>
  </options>
</cc_config>

It should straighten itself out in a couple of days then.
But the BOINC scheduler does what it wants, so don't be too concerned.

But only if the tasks have similar durations, deadlines and the projects have task availabilities and distributions rules.

But only if the tasks have similar durations, deadlines and the projects have task availabilities and distributions rules.

Indeed,
because CPDN mostly has long deadlines and long duration tasks, questions about this come up on a regular basis on the CPDN boards. I actually suspend WCG except when CPDN has no work as I am not particularly interested in fairness between projects I crunch for.