OSDev.org

In the memory management thread we've been talking a bit about Singularity, Microsoft Research's (seperate division!) managed (ie, ".NET") operating system. Personally I'm very intrigued by this and it seems some other people are as well.

For anyone who wants to get some more info or may have missed something I've collected the most important resources (all files can be downloaded for offline checking):

• Singularity homepage
• Overview presentation* (Powerpoint)
• Design motivation & short overview (PDF)
• More elaborate technical report with API listing (PDF)
• Video #1 discussing Singularity (WMV)
• Video #2 discussing Singularity (WMV)

(* the presentation contains a screenshot with kernel tracing, not really much to see though)

Personally I really enjoyed the two videos. It's clear the guys are pretty excited working on the project.

In case some people are still following this research project,
I found some new links to papers on the Singularity homepage:

• Solving the Starting Problem: Device Drivers as Self-Describing Artifacts (PDF)
• Language Support for Fast and Reliable Message-based Communication in Singularity OS (PDF)
• Sealing OS Processes to Improve Dependability and Security (PDF)
• Deconstructing Process Isolation (PDF)
• {End Bracket} Singularity

The document titles alone make me loathe Singularity. Reads like a marketing flyer. :-X

I'm not sure what you are trying to accomplish with that remark.
It sounds like you are trying to shoot the project down or invite
a flame war or something.

If you are not interested then simply don't bother with it. I know
a couple of people, myself included, are interested.

To me the titles read just like any other research paper.

Microsoft Research is not the same entity as "regular"
Microsoft. A lot of those guys have (academic) research
backgrounds and they are working with various Universities
on this project, for example.

Agreed, those titles read like any other academic paper. Might have a quick browse through there, especially the drivers ones.

Rob, shift down a gear. We're all here for the fun of it.

I have to face some snide remarks every other time I come up with "...in AmigaOS..." - because I know its internals better than those of most other OS' and have spoken about it so often.

The same is with you and Singularity - you brought it up here, and you are the one advertising it. I don't mean to offend you by calling it "advertising", it's simply what you are doing here. It's useful for people who are looking for "information entry points", and thus I welcome it. (I do! As long as it keeps contained somewhat...)

But I reserve the right to make a bit of fun on such advertising. I have done the same during the hot phase of the Exokernel hype, or whenever someone pops up (again) claiming that hard real time is a requirement for viewing MPEG videos...

I don't like research papers that try to sound too smart, personally. I positively hate marketing types who blow up simple things with so many words just to make them sound more impressive. (Meaning those who came up with those titles, not you.) One of my favourite research papers on the net is titled "recursive make considered harmful". There's a to-the-point title.

Singularity might have great concepts, and the people working on it might be brilliant. I don't know. But the titles of their research papers make my stomach boil. Perhaps I should have added more smileys like ;D or so you could have taken it more easy.

And if you think I - a humble lone coder - could aim to "shoot down" a Microsoft project, well, thanks for the compliment on the amount of influence you think I have.

Relax. Take it easy. It's a fun hobby, and a bit of banter keeps the spirit.

After a minute of thinking I can even tell you why I am allergic to smart research papers: Because I strongly believe that we don't need so many new OS concepts, but rather someone who implements the existing concepts in a useful way (i.e., having a reactive and intuitive user interface, a comfortable and consistent API etc.).

The titles of those papers read like something you are told to study in a theory centered (?!sic) course at university. *rofl*

No one can understand what osdeving is or how an os works in its innards by merely studying parts of it (scheduler, memory management, semaphores, message passing, ... and each of them isolated from the other). I've told fellow students who are on the edge of despair to give crap to the mere theory and to take the plunge - and actually code a bit. Draw stuff, design stuff and try it out for gods damned sake.

ah. I'm straddling off topic.

@Rob: take it with a grain of salt. Solar is a fine fellow. He just tends to crack quite rough jokes sometimes, but that's naught the majority of fella's here canna deal with, eh? Enjoy having a good time and cracking some jokes yourself.

Look who's talking. ;D

Solar,

You happen to have any material about Amiga OS anywhere?

I mean, would love some documentation about it's internals, but haven't found anything good (haven't searched really hard though, but the normal google with a dozen keywords).

Not that I want to recreate AOS, but love to know more about these things people always talk about..

Consider that the guys at Microsoft Research probably have to convince the MS executives of their usefulness in order to get more funding. In other words, maybe those titles weren't crafted with us in mind.

Quickly checking a few of the papers, reminded me of why I don't like "research papers" published by commercial entities.

Use 20 pages to say what could be said in half a page of an undergraduate textbook. Oh well.

Solar: thanks for your response. I'd like to think I'm a guy that
takes things easy. The reason I thought this was not cool was
that you said "loathe Singularity" due to "the document titles".

That just doesn't feel okay to me, judging a project/product by
a title of document about that project/product.

But then again maybe you meant it as a joke or something. No
harm done, it just didn't go down well for me.

Hi,



It's exteremely rare for me to read any research papers, regardless of who they're from because they always use "difficult to read" terminology - it's almost as if they're playing "scrabble" where they get extra points for longer words.

About Singularity, watch the videos! I did and I found them very interesting and easy to follow - watching normal programmers wearing jeans describe their work on a white board in a smallish back office is a lot different to any marketting material or research papers....


Cheers,

Brendan

The main reason I don't investigate stuff like this is because I trust Microsoft to have patented at least some of their "novel" ideas.