Re:Rise4fun - from Microsoft Research
by _delirium on Thursday April 12, 2012 @08:41PM
(#3834842)
Note that the projects are at wildly different levels of availability/maintenance, so if you see something interesting, google [edit: or bing :-] to find its MSR project page (not sure why those weren't linked).
At one end of the spectrum, the graph layout engine (AGL) costs $279, and its last release was Automatic Graph Layout 2007. At the other end, the Z3 theorem prover, while closed-source, has a free binary download, three MSR employees developing it, active maintenance and support, and good documentation/examples.
Re:Rise4fun - from Microsoft Research
by gosub on Friday April 13, 2012 @04:16AM
(#3835851)
I guess it could be very uncomfortable to use a theorem prover that is not open-source, especially if your work is at stake. Closed-sourceness abruptly cuts the chain of trust at Microsoft.
Re:Rise4fun - from Microsoft Research
by mturmon on Thursday April 12, 2012 @11:21PM
(#3835271)
Thanks for that, which was new to me. It seemed to do what I had in mind.
I raised this a few months ago, but commenters thought I wanted to solve a different problem, such as a tool to check a regex against a group of strings that I would supply myself:
Re:Rise4fun - from Microsoft Research
by sanxiyn on Friday April 13, 2012 @08:41AM
(#3836395)
Well, SMT solvers do prove theorems, just not kind of theorems that mathematicians care about.
Wikipedia thinks the usage is OK. "Theorem prover may refer to: Automated theorem prover, or Proof assistant, an interactive theorem prover." http://en.wikipedia.org/wiki/Theorem_prover
Re:Rise4fun - from Microsoft Research
by bjacobso on Thursday April 12, 2012 @06:38PM
(#3834500)
I hate to criticize a clearly awesome effort on MSFT's part but: why are those links not <a> tags? The result being a user is now unable to right click and open the examples in new tabs...
Re:Rise4fun - from Microsoft Research
by tluyben2 on Saturday April 14, 2012 @05:49AM
(#3840414)
I have touchdev on my lumia for quite a while as i'm interested in programming with touch devices and although it's nice, it's very unimpressive in the evolution of programming. It's still just really annoying to code on touch devices and unless someone makes something really revolutionary where you don't have to type and try to position the cursor so much, I don't see this is at all as a 'revolution'.
Re:Rise4fun - from Microsoft Research
by freehunter on Friday April 13, 2012 @08:16AM
(#3836310)
TouchDevelop is nice, but it still uses its own marketplace for applications. Hopefully with Windows 8 they'll extend it, add in a few more libraries, and allow code to be published to the official Marketplace. What's nice is that the TouchDevelop code on the unofficial marketplace is open source, you can fork the code and push your own variant back.
Re:Rise4fun - from Microsoft Research
by contextfree on Saturday April 14, 2012 @02:31AM
(#3840114)
TouchDevelop scripts can be submitted to the "real" WP Marketplace, see http://www.touchdevelop.com/help/wp7app . IIRC this is a new-ish feature, it may not have been there when last you looked.
Re:Rise4fun - from Microsoft Research
by mkramlich on Thursday April 12, 2012 @07:18PM
(#3834613)
Wanted to quickly "open in background tab" through each of them but couldn't it. Leave it to Microsoft research to not deliver web links that work right. "When you're so advanced, you can't even do the basic things!"
Re:Rise4fun - from Microsoft Research
by sauravc on Thursday April 12, 2012 @08:59PM
(#3834893)
Disclaimer: I used to work in a incubation group that was under MSR. Although there's plenty to gripe about at Microsoft, MSR is not one of them. It's a shining example of an awesome research group. (It probably employs more computer science researchers than any other company and was a great place to bounce ideas off of your peers, who are some of the smartest people in the field.) But one thing you should know, researchers aren't great web developers. Just look at any top computer science professor's personal websites. Judging a project's worthiness by its website is exactly like judging a book by its cover. It's pointlessly idiotic. At least balance it out with a comment about the projects on the page.
Re:Rise4fun - from Microsoft Research
by knieveltech on Thursday April 12, 2012 @10:27PM
(#3835132)
"Judging a project's worthiness by its website is exactly like judging a book by its cover. It's pointlessly idiotic."
A valid point, but there's another side floating around here unmentioned. I've lost count of the times I've had CS types sneer at my chosen profession (web development) as "child's play" and "not real programming". And yet for a task that is apparently infantile it's instructive to see how frequently it's done poorly by so-called "real" programmers.
Re:Rise4fun - from Microsoft Research
by latch on Thursday April 12, 2012 @08:28PM
(#3834820)
I think it's constructive to tell them their site is broken (which I also agree that it is[1]). However, you are also right that your parent was unnecessarily condescending.
[1] There are billions of links which are just plain <a href="X"></a>, so the fraction of the ones that do something weird (like this, or target="_blank") really break the browsing experience for people.
Re:Rise4fun - from Microsoft Research
by eta_carinae on Thursday April 12, 2012 @09:24PM
(#3834969)
I'm a pretty big supporter of Microsoft overall but I came here to post just this. There are a bunch of projects on this page and I wanted to open all of them in background tabs.
I can't.
What's the justification for not using plain links for this page, seriously?
Re:Rise4fun - from Microsoft Research
by sauravc on Thursday April 12, 2012 @09:08PM
(#3834917)
Rise4Fail? Really? So basically since the book's cover is ugly the story must be a fail? Obviously that's the case. Brilliant researchers always have great looking websites, right?
Re:Rise4fun - from Microsoft Research
by geuis on Thursday April 12, 2012 @11:27PM
(#3835287)
I commented on the unusability of the site that was linked to. Mr. Knuth's site is perfectly readable because he's using regular html which renders just fine on a modern mobile browser. Meanwhile I have no idea what is going on on the parent page, except that it's unreadable. Another person shared a similar screenshot in another comment.
You may dislike my snarky little hashtag on Twitter and that's fine. I was not being snarky here on HN in my comment. I was making a perfectly valid criticism about usability.
I attempted to parse that as a sentence before realizing it was a navigation bar
I thought it was something generated by a poorly constructed markov chain.
At one end of the spectrum, the graph layout engine (AGL) costs $279, and its last release was Automatic Graph Layout 2007. At the other end, the Z3 theorem prover, while closed-source, has a free binary download, three MSR employees developing it, active maintenance and support, and good documentation/examples.
Of course, L is usually infinite, so it's easy to generate strings in an unhelpful way.
For instance, the phone number one:
http://www.rise4fun.com/Rex/phone
and the URL one:
http://www.rise4fun.com/Rex/weburl
You can modify the given regex, click "ask rex", and wait a while, and it will return two patterns from the regex. Just two...
I raised this a few months ago, but commenters thought I wanted to solve a different problem, such as a tool to check a regex against a group of strings that I would supply myself:
http://news.ycombinator.com/item?id=3297788
http://www.smtcomp.org/2011/
Wikipedia thinks the usage is OK. "Theorem prover may refer to: Automated theorem prover, or Proof assistant, an interactive theorem prover." http://en.wikipedia.org/wiki/Theorem_prover
What I would do if I wore Steve Ballmer's shoes for a day: http://www.rayhano.com/post/11818893068/microsoft-research-s...
nice, incentive to write code on mobiles finally. I guess the mobile revolution will came from microsoft after all. oh the irony.
A valid point, but there's another side floating around here unmentioned. I've lost count of the times I've had CS types sneer at my chosen profession (web development) as "child's play" and "not real programming". And yet for a task that is apparently infantile it's instructive to see how frequently it's done poorly by so-called "real" programmers.
The projects definitely look cool though. Looking at Koka now, in the original tab...
[1] There are billions of links which are just plain <a href="X"></a>, so the fraction of the ones that do something weird (like this, or target="_blank") really break the browsing experience for people.
I have to report that I tried to mess around with the AGL example and got a perpetual loading screen. The concept seems good; the execution, lacking.
Maybe this is not supposed to be public?
I can't.
What's the justification for not using plain links for this page, seriously?
http://t.co/Belra6Lw
I literally have no idea what this page is supposed to be, given how absolutely horrendous it renders.
http://www-cs-faculty.stanford.edu/~knuth/
You may dislike my snarky little hashtag on Twitter and that's fine. I was not being snarky here on HN in my comment. I was making a perfectly valid criticism about usability.