Follow Slashdot blog updates by subscribing to our blog RSS feed


Forgot your password?
What's the story with these ads on Slashdot? Check out our new blog post to find out. ×

Comment Re: Programming (Score 1) 601

I'm going to have to say web "coding" is not coding/programming at all. Which makes the article pointless (though we all knew that). Editing HTML/XML (or IMO even most Javascript) is about as much programming as learning to add oil to your car (Javascript is more like learning to change your own oil). Neither one makes you a mechanic.

Comment Re:Hard to believe (Score 1) 172

t's also not pointless for software decoders such as VLC that have access to a bit more memory and CPU capability to deal with it.

If you are talking flexibility of decoders and extra CPU, why make a non-compatible file based on a decade+ old codec when you could just re-encode to H.265? Same with JPEG92 vs JPEG2000.

Along those lines it doesn't seem that long ago that arguments about using floating point in mp3 decoding was seen as a flaw.

Totally different issue, since it wasn't about making the codec non-compatible, but how it's decoded.

So, not entirely pointless, but not nearly as interesting as the article pretends it is, which is typical of business journalists who don't really understand tech. Many of the "inefficiencies" have already been solved with more recent codecs. Retrofitting things like variable macroblock sizes and alternate compression strategies onto old formats is not particularly revolutionary...

Comment Re:This isn't really a new thing. (Score 1) 172

Postprocessing software like Beamr (look it up yourself...) can often do even better for video. Basically the H.264 codecs are fairly conservative on their quantizers, with a minimum that's way above what they could get away with. Way better off throwing away useless data than figuring out how to compress it.

Comment Re:Hard to believe (Score 1) 172

And they probably say mobile-encoded for a reason, simple encoders have to work on low power and in real time,

Actually, the encoders are rarely limited by power or CPU cycles. The decoders are, but the great thing about lossy encoding like JPEG/H.264/H.265 is the encoders can continually be improved without affecting the decoders.

That said, the reason this article is pointless is you can't USE the results - it breaks H.264 standards so HW decoders can't handle it, and no one wants to decode some proprietary format on the fly to stream to standard H.264 decoders...

Comment Re:Hard to believe (Score 1) 172

The goal of H.264 and JPEG isn't minimum file size at all costs. It's also not encryption. Your premise is wrong, and even old tech can compress this stuff further than it may already be.

True, but that's obvious to you and me - which does reinforce the point that the article & Dropbox "innovation" is pretty stupid.

Not to mention JPEG and H.264 are old news - if you want to compare "new" development JPEG2000 and H.265 are the benchmarks...

Comment Re:naysayers are missing the point (Score 1) 172

Except unless standard DECODERS can handle them it's fairly useless in practice.

From what I can tell from the source & description posted it does NOT conform to H.264, so what's the point? SOMETHING has to decode it, and it's clearly not going to be standard hardware decoders. So it's useless as CDN storage. Same applies to PNGs for most usage.

And besides, H.265 implements everything they did and MUCH more. And if you want even further lossless compression that humans can't notice there are proprietary solutions like Beamr that can get you even better compression.

Comment Re:Cost (Score 1) 176

If you are living in a very rural town, maybe (in which case, the town should deal with it, but it's SO far from a global problem - either a totally isolated issue or, in the Bay Area, likely an utter 1% first world problem). If you are saying horse manure is a problem in SF, well, I call horseshit on that...

Comment Re:Simple solution (Score 5, Insightful) 176

That's like charging a "speeding tax" to everyone on the road and then letting people drive as fast as they want.

It's a minority of the dog owners in most towns that don't pick up after their dogs. Charging everyone for a few shitheads violating the law is bad policy, no matter what the offense is.

Comment Re:Sad Birds (Score 2) 109

If you are working "so many hours", you better be getting something significant for it. Because all of those developers you can't find clearly are, or you'd be hiring them. So if you are not happy with your job or hours and think you are worth more - stop whining, the jobs are out there!

Also: if you work at a startup, have tons of money in the bank and you can't hire, either the work is horrible or you just aren't offering enough. I hear this all the time: "all of the good candidates are going somewhere else!" Well, if you got outbid, no shit they took a better offer. Try harder next time. And if the company/job sucks, you have to pay even MORE to get someone to do it.

If you have 3 times the open positions as employees, how about paying everyone 50% more? Then the current employees and the new hires will both be happy. Unless you don't think you or your coworkers are worth it?

Comment Re:Sad Birds (Score 1) 109

Yup, it's a basic matter of a tiny indie game company that inexplicably had $1B of unexpected revenue dropped in their lap all of a sudden.

They did what many companies, game dev or not would do, and tried to expand as fast as possible. The difference is game companies (with maybe Blizzard/WoW as an exception) can rarely sustain that revenue to the next year without another massive hit. And I'm sure as you know, there are very few left (maybe Rockstar and Bethesda?) who have managed to create that string of massive hits to justify hiring without even having a semi-guaranteed new hit in place...

I bet some big publisher offered them a giant pile of money in the last few years and they rejected it. And I bet they are now kicking themselves (as compared to Notch, who now owns the most expensive house in Beverly Hills...)

We declare the names of all variables and functions. Yet the Tao has no type specifier.