=> More informations about this toot | View the thread
=> More informations about this toot | View the thread
The closest thing I know of is something like ~/.cache but that doesn't take into account build artifacts and such although maybe that's something build systems should change.
=> More informations about this toot | View the thread
I half-seriously want something that could be described as an oodkiller. More specifically, I wish more "disk space clean-up" applications had better metadata semantics for all the random crap on your file system. I wonder if there's some kind of standardized opt-in scheme that applications could use to register or tag intermediate build artifacts, cache files, stuff that could be on-demand reinstalled, etc.
=> More informations about this toot | View the thread
Unfortunately I have to say this new NG2 Black in anything like its current state is much worse than even NG2 Sigma. I played through the first few levels of both NG2 Black and Sigma back to back and while Black put the gore back and there's a few more enemy spawns it's still nowhere near the original NG2. And the performance is absolutely atrocious, frame pacing issues almost makes it unplayable even at low quality settings. Input lag is through the roof, too, and it seems to drop inputs.
=> More informations about this toot | View the thread
The NG2 remaster mostly just made me upset and now I'm casually looking at prices of used Xbox Series S for running NGB and NG2 in backcompat. Hard to justify but it would almost be worth it for those two.
=> More informations about this toot | View the thread
Speaking of recursive filtering, this radial blur technique by Rune Vendler made a big impression on me at the time. It was the workhorse behind the radial blurs in all those 64k intros by Fudge from the mid to late 1990s: https://gist.github.com/pervognsen/c57b9c25c884cf67f5933ac1aa058052.
Example: https://www.youtube.com/watch?v=IgF8KRNi2gA
=> More informations about this toot | View the thread
Blast from the past: https://www.youtube.com/watch?v=0oOCobldnmY. I remember those slides and sample programs by Kawase were pretty influential at the time. E.g. "Kawase blur" became synonymous for a while with the recursive filtering technique where you do a multi-pass blur and double the tap spacing each pass, e.g. the product formula (1 + z)(1 + z^2)(1 + z^4)... = 1 + z + z^2 + z^3 + ... for the geometric series. Also an early example of a game using pre-convolved light probes for IBL.
=> More informations about this toot | View the thread
The worst and not that rare failure mode is when I end up having to truncate at the professor's home page because that's the longest prefix that can be truncated without a 404 and then you have to do a deep forward search from there to find the siblings.
=> More informations about this toot | View the thread
I run into this surprisingly often and it's become a pet peeve of mine: if you have a URL structure (e.g. for a course website) please make it easy to truncate the URL at every separator in a way that will get me to sibling resources. E.g. I'll often find a link to some lecture notes or a paper for a course website and want to see any other lecture notes for that course.
=> More informations about this toot | View the thread
=> More informations about this toot | View the thread
=> More informations about this toot | View the thread
=> More informations about this toot | View the thread
Same-day announcement and release of a UE5 remake of Ninja Gaiden 2 was pretty unexpected. https://www.youtube.com/watch?v=r-UB8SOyru0. Cue the NG2 Sigma haters already complaining in the Steam reviews that this is more of a Sigma remake.
Platinum is also making Ninja Gaiden 4, which looks nothing like Ninja Gaiden and given Platinum's long string of failures I think expectations should (unfortunately) be set very low.
=> More informations about this toot | View the thread
=> More informations about this toot | View the thread
By the way, I like the space-time variant of the standard proof that van Emde Boas gives in https://people.irisa.fr/Francois.Schwarzentruber/sif2_thx/articles/tiling.pdf. "Basing the proof on the space-time diagram, uniqueness of the head position is a direct consequence of local consistency throughout the diagram and does not have to be enforced explicitly."
=> More informations about this toot | View the thread
(If the tape head starts in a fixed initial position then within time T you can only access tape cells within distance T of that initial position, so there's no need to encode instances of the transition relation corresponding to machine states beyond the "light cone" emanating from the initial position.)
=> More informations about this toot | View the thread
To do the first step (encode the transition relation in propositional logic) you need to unroll in space, not just in time. In the case of Cook-Levin that relies on the usual argument that TIME(f(n)) is a subset of SPACE(f(n)) since the Turing machine's tape head starts in a fixed position, so the nondeterminism is only for the transitions. In model checking I think they usually allow an arbitrary initial-state predicate, so I guess for BMC you'd have to constrain that as well.
=> More informations about this toot | View the thread
I hadn't really thought about it in those terms, but the implementation of bounded model checking is very similar to the proof of the Cook-Levin theorem. You encode the nondeterministic transition relation of your model of computation as a formula in propositional logic, unroll it to some fixed depth (given by the polynomial-time bound in the case of Cook-Levin), and check for satisfiability of the unrolled transition relation formulas and an auxiliary formula.
=> More informations about this toot | View the thread
=> More informations about this toot | View the thread
=> This profile without reblog | Go to pervognsen@mastodon.social account This content has been proxied by September (3851b).Proxy Information
text/gemini