Forgot your password?

Comment: Re:OMFG compile! (Score 1) 109

by tlambert (#46821453) Attached to: WRT54G Successor Falls Flat On Promises

yeah so linksys developed the chips and stuff? yeah right(1).

(2) other manufacturers buy from the same chip manufacturers and get the same cookie cutter drivers under the same cookie cutter nda.

3,4,5 doesn't stop others from doing so.

Making the driver proprietary and licensed only for use with the OpenWRT hardware most certainly does prevent 3,4,5 from being leveraged by another vendor. If all I have to do is copy your commodity chip choices and a lot of your interconnect design, you have R&D costs to recover, and I don't = I put you out of business, unless you have brand loyalty above and beyond the price point.

Can you say, with a straight face, that another product that could run the exact same software load and work the exact same way wouldn't render the OpenWRT router fungible, and therefore the only thing that would matter to consumers was the price point? I can pretty much guarantee you can't do that, even if every person who was in the market for an OpenWRT-style device pledged to buy theirs instead of a cheaper one.

You also have failed to address the SDR issue of FCC/CCITT licensing the combination of hardware and firmware as a unit, and revoking licensing for hardware that can be used with a different firmware and/or firmware that can be used with different hardware, because they don't want to have to deal with malicious radio loads screaming over top of military radio bands because the hardware is capable, but it's only the firmware which limits the ability of someone to do the nasty.

Comment: Re:OMFG compile! (Score 2) 109

by tlambert (#46821439) Attached to: WRT54G Successor Falls Flat On Promises

As someone who has worked on a Linux-based embedded system, and had to cross-compile to do it... dude, Linux cross-compilation sucks, and there's almost universal pushback from everyone wo deals with Linux build systems, from Debian to Red Hat, and beyond, to any attempts to make it better.

Did you try OpenEmbedded / the Yocto Project? It takes away pretty much all of the pain of cross-compilation. Most of our users seem pretty happy with it.

Yocto has a different goal than cross-building a standard Linux distribution, along with some components. I was specifically involved in ChromeOS, and the cross-build wasn't there fore something as large as a complete Debian distribution.

I think the big problem with Yocto and OpenEmbedded (or ChromeOS) is that it assumes a Linux host environment, and acess through the host environment to package management tools.

I admit that there was a lot of intrinsic bias because of the team's history towards a Debian-based system, but our desktops were a Debian-style environment as well (Ubuntu), and the common implementation was to chroot into a more or less "pure" Debian build environment on the desktop, and then from that, chroot into a cross-build environment basically identical to the first chroot environment, and from that do the cross-build, including installing build products from the second chroot into the build environment for the target, and using them.

Neither Yocto nor OpenEmbedded addresses this issue adequately -- while Yocto did something to eliminate about half the hassle when Richard Purdie did his patch set last October, I don't think it was enough to get to the point where you could base a ChromeOS on it; you could use it for a single embedded device, and, with a lot of work, a number of packages on the device, but clearly nothing like all the software (which I freely admit - it's too much code) needed to do the full product, or do it in a way that was convincing enough that the additional work warranted abandoning a working (non-cross) environment.

Comment: Re:OMFG compile! (Score 3, Insightful) 109

by tlambert (#46821209) Attached to: WRT54G Successor Falls Flat On Promises

You know the man pages are the manual right?

How about you bother to learn something instead of coasting on the work of others for a decade then complaining things don't fulfil your every need after you've contributed exactly bugger all.

As someone who has worked on a Linux-based embedded system, and had to cross-compile to do it... dude, Linux cross-compilation sucks, and there's almost universal pushback from everyone wo deals with Linux build systems, from Debian to Red Hat, and beyond, to any attempts to make it better.

IMO, you should be able to download and install OpenFriggingSolaris on a SPARC system, and cross-compile Linux for ARM, Alpha, and Intel on the damn thing, without having to have some dumb-ass chroot environment because someone is too stupid to deal with include paths, library paths, and source paths correctly, and because the build process somehow thinks it's an OK thing to use build products created during the build process as part of subsequent build steps. I mean, how incredibly, obviously stupid is it to use intermediate build products as part of your build process, unless they are targeted solely at your host environment, and never mirrored into your target build product area (oh yea, a working "DESTDIR=" would be kinda helpful here, too...).

The whole idea that you can have dependencies that reference files in the host environment other than those on a mounted read-only source partition, and that "retry" package builds each time because the build system is too stupid to figure out missing dependencies is terrifically annoying.

Comment: Re:OMFG compile! (Score 0) 109

by tlambert (#46821189) Attached to: WRT54G Successor Falls Flat On Promises

RTFA. The patches are a mess,

Yeah, not seeing this one as a problem; Open Source projects have no problem supporting hardware that the manufacturer would rather they didn't support, often over the manufacturers objections, but when it comes to hardware specifically built on behalf of the Open Source project, all of the sudden it's now the companies job, rather than the Open Source project's job, to pee on the patches until they smell like the projects leaders peed on them, so that there are no changes required to be able to use them.

This seems really similar to Samsung releasing code with "board" support for some hardware, and then some maintainer getting all pissy that they didn't write the code the same the maintainer would have, had the maintainer had the time, but the maintainer doesn't have the time, but won't integrate the patches anyway because they aren't done the same way they would have been done, had the maintainer done them, but the maintainer won't do them.

Either bitch when they don't obey the GPL and provide their code, or take their code when it's provided and say "thank you", but don't bitch when they hand you code, and you don't want to do the work to integrate it into your moving target of a project. Thanks.

don't compile cleanly and the wireless driver is missing. Rendering it an expensive paperweight.

It's not entirely a paperweight, but they've acknowledged that the code, as supplied, lacks wireless driver support, and that they need to sanitize the code and break it along interface boundaries so that it can be a binary driver module.

Again, I think the "problem" isn't so much that the module wasn't supplied immediately, instead of just being promised, but that it means they aren't going to get the source code for the module itself. A lot of Open Source projects like to try to force hardware vendors to give up what the hardware vendors consider their "keys to the kingdom", and will go so far as to design system interfaces which aren't usable unless you have GPL'ed code in your driver, making your driver GPL'ed, meaning that they can demand source code.

As far as SDR - Software Defined Radio - such as that used in WiFi and cellular radio parts firmware is concerned, those guys can piss up a rope. Specifically, if the source were made available in a way that could be utilized the way the Open Source people want it to be able to be utilized, which would mean:

(1) Other vendors could just copy the register interfaces and use the same driver, without having to do hardware design work
(2) Other vendors could thus undercut the prices by the amortized R&D costs (i.e. the hardware would be commoditized)
(3) The driver work would effectively not be a recoverable cost at the commodity price point
(4) They lose their FCC certification for the part
(5) They can't sell in the U.S., France, the U.K., Japan, and other countries that license hardware/firmware as a single lump

So... piss up a rope; be happy with the forthcoming binary-only driver blob, and be happy it's been promised at all so that you can dick around with the way the rest of the system works to your hearts content. That's all you're going to get for economic reasons, unless you get together as a group and buy out their R&D costs, and buy out their first mover advantage.

Otherwise, if you can live with the limitations, hold your damn horses, and wait to buy the router, which is generally not hardware available anyway.

Comment: Re:Obligatory (Score 2) 307

It'd certainly be a good border security method against Mexicans. In fact, they could start by just targetting drug runners and practically solve the drug problem overnight. Drug dealers cost America more money and kill more americans than terrorism by about 100000x

When did drug smuggling become a capital crime?

I'm pretty sure it happened about the time interdicting drug smuggling involved risk of death.

And I'm pretty sure *that* happened about the time that being a successfully interdicted drug smuggler carried huge penalties, including life in prison.

And I'm pretty sure *that* happened when people starting smuggling huge rather than trivial amounts of drugs.

And I'm pretty sure *that* happened about the time the economic incentives for smuggling became so large.

And I'm pretty sure *that* happened about the time we announced a "war on drugs".

And I'm pretty sure *that* happened about the time the CIA started using Heroin from Southeast Asia to fund the covert "War on Communism".

And I'm pretty sure *that* happened as soon as we cut off their legitimate sources of funding, but kept their goals, tasks, and targets the same.

Comment: Re:"no indication ... site has been compromised" (Score 1) 80

by tlambert (#46803397) Attached to: Preventative Treatment For Heartbleed On

The site doesn't have any medical information at all. That's one of the advantages of outlawing the "pre-existing condition" scam - you no longer have to tell insurers your medical history to buy insurance.

No, you still have to tell them; that provision of ACA doesn't occur until the end of this year, after you are already enrolled (by which time, it's too late). Until then, they have to let you enroll, they don't, however, have to charge you a reasonable monthly rate if you have a pre-existing condition. They said they had to let you buy it, not that it wouldn't be expensive. That one of the reasons the first 'A' in 'ACA' is a bit misleading.

Comment: Re:"no indication ... site has been compromised" (Score 2) 80

by tlambert (#46797835) Attached to: Preventative Treatment For Heartbleed On

If only it could have been prevented via a cheap, preventive program, instead of costing so much later! I know! We should lobby them to create a new agency, one tasked with the security of the nation, and when they knew about risks like this, why, they could step in and ensure that no one would unwittingly deploy vulnerable systems in the first place!

Perhaps we could call them the Responsible Agency for Intelligently Securing the Interests of the Nation... R.A.I.S.I.N., for short... or National Organization Securing You... N.O.S.Y. for short... I'm still working on the name.

We could even nominate someone to put in charge of making sure they are doing the job they are supposed to be doing, a kind of Special National Operations Watch Director Executive Nominee... Haven't decided what to call that one yet, either...

Comment: "no indication ... site has been compromised" (Score 4, Funny) 80

by tlambert (#46797719) Attached to: Preventative Treatment For Heartbleed On

"no indication ... site has been compromised"

I believe them.

What possible motive would a hacker have for targeting a site containing social security, tax, medical, personal, and financial information?

I'm sure it's all perfectly secure.

Just in case, though, you should probably change your one-factor authentication token so that the next time your "keep me logged in" cookie expires, it's hard to remember.

Comment: Re:"...who exactly is the H1-B police..." (Score 1) 220

by tlambert (#46797677) Attached to: California Utility May Replace IT Workers with H-1B Workers

Well there's an authority to base your response upon! I especially like the links to payday loans and making sure your H1B sponsor treats you properly. I missed the part where it actually backs up a single thing you assert since the press release it references is not linked.

Government press release you could have googled yourself. Feel free to continue whining that nothing is ever enforced in this area of law.

Comment: "...who exactly is the H1-B police..." (Score 2) 220

by tlambert (#46793523) Attached to: California Utility May Replace IT Workers with H-1B Workers

And who exactly is the H1-B police who come arrest the violators?

That would be:

= U.S. Immigration and Customs Enforcement (ICE)
= U.S. Citizenship and Immigration Services - Fraud Detection and National Security Division (FDNS)
= U.S. Department of Labor - Office of Inspector General
= U.S. Postal Inspection Service (USPIS)
= U.S. Department of State
= U.S. Attorney’s Office for the Southern District of Iowa

At least that's who it was for this case:

So perhaps you are an idiot for implying that these laws are unenforced and unpoliced, and it's a scaremongering tactic which actually has very little to do with the offshoring indicated by the original article, which in turn has very little to do with H1-B's at all, since off shore workers are in other countries, and don't require H1-B visas to be employed by a U.S. company, if they never leave their home country.

Comment: Re:Profits (Score 1) 336

by tlambert (#46790797) Attached to: Detroit: America's Next Tech Boomtown

Given that Ford earned $7.2 Billion in net income in 2013 and GM made a $3.8 billion profit over the same period I think GM and Ford will be very surprised to hear that they cannot make cars in the US profitably since most of their profit comes from US operations.

They'd only be surprised if you told them they'd be doing it in Detroit, instead of non-union plants in other U.S. states:

You don't need to expand factories to make the efficient.

Correct. You just need to reduce the number of employees to increase the profit per employee, which is something you can do with automation, and.or lower wages, which is not something you can do in Michigan.

Comment: Re:FLYOVER (Score 1) 336

by tlambert (#46790703) Attached to: Detroit: America's Next Tech Boomtown

If you're interested in high tech manufacturing with a skilled workforce, it would be hard to find a better place than the automation alley counties. What you'll spend in wages will be more than made up in productivity. And you won't be spending a fortune in recruiting costs. If you build a factory your staffing problem won't be finding qualified workers, engineers or tradesmen, but getting a big enough HR department to hire them.

The reason all but one automotive assembly line has pulled out of Detroit is that the unions wouldn't allow that much automation, or you were "allowed" to have it, but you had to still hire the same number and type of workers to satisfy the contracts, so it didn't do crap to change your value to unit labor cost ratio.

You are an absolute idiot if you locate a manufacturing facility in a state where the unions are in charge of whether or not you get labor, and you can't push costs down by automation.

Most blue collar jobs have migrated outside the U.S. due to inflated labor costs relative to value produced. It has dick all to do with what a living wage is or isn't, and *absolutely everything* to do with value produced per unit labor cost. Most U.S. auto manufacturing that still exists in the U.S. at all is in non-union states, in non-union shops.

As Steve Jobs said, "Those jobs are gone, and they're not coming back". Near the end, before they sold it to Canon, the NeXT factory producing laser printers required exactly two (2) full time workers to operate the entire factory.

Comment: Re:Almost all router bandwidth management is shit. (Score 2) 104

by tlambert (#46786077) Attached to: Ask Slashdot: Which Router Firmware For Bandwidth Management?

OK, as someone who has been trying different methods of QoS over the past years, with varying levels of success, mainly to have my VoIP phone rock solid over DSL, I'm very interested in what you're saying.

Is there a reason this approach hasn't been implemented yet? Does it break something? If my router is lying to one my upstream router about its TCP window size, wouldn't that impact both the FTP and video stream?

You lie about the window size on a per connection basis, so no, since it's not a global policy, it's a resource policy by application, and potentially by port/IP tuple, so it's not a problem. The point is to keep the upstream router packet buffers relatively empty so that the packets you want don't have to be RED-queued. Nothing breaks because of it.

It generally won't work, unless everyone "plays fair", and the port overcommit ratio for upstream vs. downstream bandwidth is relatively low. As the downstream data rate increases to approach the upstream data rate, the technique loses value, unless you get rid of overcommit, or do it on a per-customer "flow" basis (as opposed to a per virtual circuit "flow" basis) within the upstream router itself, or move to a "resource container" or similar approach for buffer ratio allocation in the upstream router.

So in theory, Comcast (as an example) could do it if they made everyone use the router they supplied, and their routers all participates in limiting upstream buffer impact.

Maybe the next time they replace everyone's cable modems, they'll bother to do it?

Without the deployed infrastructure, it's easier to RED-queue and just intentionally drop packets, forcing a client to request a retransmit as a means of source-quenching traffic. This wastes a lot of buffers, but they probabilistically get through, and for streaming video, that's good enough if there's a lot of client overbuffering going on before playback starts (JWZPlayer, for example, is a common player used for pirated content that will habitually under-buffer so intentional drops tend to make it choppy).

For VOIP, unfortunately, forced retransmit causes things to just typically suck, unless you use a sideband protocol instead, where the router at the one hop upstream peer agrees to reserve buffers for specifically that traffic. This is why Skype is terrible, but your phone calls over your wall jacks which are actually wired to the same packet interface instead of a POTS line are practically as good as a land line or cell phone.

Google hangouts tend to get away with it because they are predominantly broadcast, and are either "gossip"-based CSMA/CD (ALOHA style) networks between participants (i.e. people talk over each other, or wait until the other end is done before talking themselves). It means they tolerate large latencies in which 1:1 VOIP/Skype connections won't. They can be a bit of a PITA for conference calls because of that (Google uses it internally, and gets away with it, but mostly because Google has its own, parallel Internet, including transoceanic fibers), but if Google employees never see the problem, they never fix the problem. Same way any company that assumes local-equivalent bandwidth works as well for their customers as it does for them (free hint to Microsoft inre: Office 386 there).

Numeric stability is probably not all that important when you're guessing.