Forgot your password?
typodupeerror

Comment: Re:It's sad (Score 1) 402

Is samsung that bad?

Last Samsung phone I has was an unlocked Galaxy Ace. It came with a bunch of useful software, and I was surprised how much I had to install when I got a Nexus phone. The stock android seems to be missing among others, some very old fashined features which have been standard since the 80s, or at least the early 90s.

Comment: Re:good (Score 1) 402

The sole annoying thing about it was that the Chinese vendor's taste in wallpapers doesn't necessarily match Western tastes (it wasn't quite Hello Kitty, but close), but that was quickly fixed.

I'm western and I set my wall paper to something which isn't quite Hello Kitty but is close.

I get irked by the forced soullessly bland corporate inoffensiveness which seems to be popular round these parts at the moment. I'm actually heartened to hear that that phone vendor has not also suffered from that particular disease.

Comment: Re:Bash is a very crappy programming language. (Score 2) 323

by serviscope_minor (#48018807) Attached to: Bash To Require Further Patching, As More Shellshock Holes Found

Even "sed" (the text filtering utility) is a programming language.

Indeed it is. It is a fantastic Turing tarpit. It was never ever ever designed to be turing complete and the feature set is geared towards doing nothing but batch search and replace.

Naturally this means that people like to program interesting things in it for fun. Some of the most impressive are implementations of dc and sokoban.

Comment: Re:Obj-C (Score 1) 310

Yes if fragmentation happens then GC wins.

However, unless you track writes, you do have to keep rescanning the entire heap (or whichever fraction is pointer data).

Many languages with GC are capable of stack allocation in some way; other languages simply have GC that is so fast that they dispense with a stack altogether.

This I disagree on: nothing beats free. The jvm has one of the best GCs in existence yet one of the analeses is escape analysis which tries to replace GC allocations with stack ones since the latter are faster.

Comment: Re:Boeing bought more politicians. (Score 1) 125

Then give the launch to the lowest bidder.

No that's a totally dumbass idea because it's so easily gamed. I can always undercut the competition. Sure the launch will fail, and they might even try to get their money back. But by that stage, it's all been spent so the company will simply fold.

Going for the lowest bidder no matter what is a great way to get screwed over by someone unscrupulous and frankly very few companies operate such a policy either. I actually had some building work done a while back and ended up going for the highes bidder. They provided a detailed cost breakdown and I could verify the costs of materials etc because I could make the same measurements, look up online and get about the same numbers myself give or take.

There is no way the other company could have done the job for the price they quoted, and certainly not if they wanted to make any profit at all. Well, I contracted out to them and they basically did it quickly, cleanly and efficiently and kepth the neighbours happily. I later heard from the people who recommended the cheap builders. The cheap ones ended up bailing halfway through another job, presumably because they couldn't finish it for the price.

So, in conclusion, going for the lowest bidder always is a terrible idea.

Comment: Re:Obj-C (Score 1) 310

I think you misunderstand my point.

In order for memory to be reused, the collector has to run, whereas with reference counting the memory is available instantly (plus the time for the slower allocation routines to run).

There's a number of papers which have evaluated the performance of GC against optimal memory management (the one that frees memory as soon as it is not needed). For about 2x the usage, the GC doesn't have much penalty, so the advantages of faster allocation etc of GC start to show. However it asymptotes towards infinite CPU as the memory usage approaches the optimal bound.

GC requires the optimal amount of memory multiplied by some constant to be efficient, whereas RC does not. In other words, RC uses the memory more efficiently. If you have memory to spare, GC can win[*], but if memory gets tight, RC will win.

The other point is that languages that do mamagement with RC typically include C, C++ and ObjectiveC all of which are also capable of doing stack allocation, which is faster than GC, so that complicates matters a bit.

Comment: Re:How it happened? Easy: PATENTS expired. (Score 1) 69

by serviscope_minor (#48008275) Attached to: How 3D Printers Went Mainstream After Decades In Obscurity

Your analogy is deeper than you realise.

We have the hit-a-button-and-get-a-perfect-print-every-time systems, however much like the computing scene of the 1980's (I'd say the 3D printers have now moved from 1970s to 1980s in computer terms) those machines are obscenely expensive to the point where no hobbyist, home or even small business user (unless they REALLY have to) considers them as a viable option.

Comment: Re:3D plotter (Score 1) 69

by serviscope_minor (#48008263) Attached to: How 3D Printers Went Mainstream After Decades In Obscurity

Still a long way to go - they are still just a toy.

I'm always amazed at how dismissive people are of other's work. Yes they hae a long way to go but plenty of people use the for actual stuff. They are a long way away from being mere toys.

The process has a long way to go - plastic is a nicer material than some home-brew thing could made, but it's still having problems. Cleaning supports and struts is a pain - I understand if you have a completely "floating" support that they are necessary but in, say, a teacup the whole thing is joined to every other point so there's no real reason to require supports. Moving up AND down a level and being able to orient the head would help a lot here and solve some other problems.

The problem is long solved, it's just a matter of expense. You can print support materials out of something different if you have a more expensive dual head printer. Then you can use a solvent or hydrolyser to remove the support material. On cheap (not strarasis) machines, it's still relatively new and experimental. It works pretty well even now however.

If you print circles, inevitably you have to adjust the print movement or else you end up with a "seam" where the head completes the circles and moves up a level.

I've never seen this: I think slic3r randomizes the start position by default. If not, there's a check box setting you're missing.

you have to wonder why the speed isn't anywhere near the best plotters as were around 20-30 years ago

Ink flows more freely than plastic. The printers can move much, MUCH faster than the print speed. They typically do so when positioning the head. However, if you try to extrude the plastic too fast it doesn't stick properly and becomes inaccurate.

Comment: Re:Obj-C (Score 1) 310

It is inferior in just about every way (runtime overhead, latency, memory utilization) to a good garbage collector.

It's certainly not inferior in terms of memory utilization because it's precise in that it deallocates as soon as the last object pointing to it disappears, whereas in normal GC, you have to wait for the allocator to run.

1 Mole = 25 Cagey Bees

Working...