* Posts by bazza

1950 posts • joined 23 Apr 2008

Google licks its lips at sight of Qualcomm's 64-bit server ARM chips

bazza
Silver badge

Intel and Microsoft need to look out.

Data centres could be on the cusp of switching to ARM very rapidly. If you can save a bunch of cash in electricity bills by swapping kit, you would. Google may be about to.

If data centres do change, Intel won't be at the party and Windows Server(ARM) doesn't exist.

18
2

Samsung trolls Google, adds adblockers to phones

bazza
Silver badge

Re: "Surprising findings"

Online advertising in the UK costs every wage earner here about £230 per year (cost of goods in the shops to pay for the advertising), whether or not you actually have a smart phone. I'd say a £10/month fee for online services would represent extremely good value, provided that meant no more online ads.

2
0
bazza
Silver badge

The huge mass of Javascript that seems to accompany a lot of websites these days is beginning to make browsing slow even on a good desktop. I can't believe how many CPU cycles are being used these days to render and display really quite simple stuff.

Layers and layers of libraries, scripting languages etc. might be fast and convenient for developers, but they're wrecking the end user experience.

9
0

Chip company FTDI accused of bricking counterfeits again

bazza
Silver badge

Re: Goodbye FTDI

This will backfire nastily on FTDI, since they are deliberately screwing their customers, the purchasers of the chips - who buy the cheapest parts they can, obviously, often via brokers. i.e. they are making their problem into a much bigger and expensive problem for their customers.

No they're not. They're not FTDI's customers at all. The people being screwed have bought fake parts;. They're getting screwed by rip off merchants illegally using FTDI's brand, logo an reputation and are so lazy they can't even be bothered to rip off FTDI's silicon properly (which is why FTDI can pu ou booby-trapped drivers)

FTDI would like them to be their customers, but are not allowed to trade on a fair market because too many people don't care where they buy from or whether it is genuine so long as it is cheap.

6
4
bazza
Silver badge

Re: Goodbye FTDI

The lazy ones are FTDI because instead of acting to clean up supply chains and chase counterfeiters they punish innocent consumers caught in the crossfire who have no way to stop this from happening other than to avoid FTDI devices altogether.

The people getting stung are not FTDI customers. FTDI's supply chain is clean. If someone chooses to buy from someone claiming to be an FTDI seller without checking, that hardly FTDI's problem.

4
6
bazza
Silver badge

But how does one know that they have a product with a counterfeit chip in it?

Try the latest driver!

I suspect that returning it the store/sales point or manufacturer will have no impact.

Depends where you live. Here in the UK If you've bought with the credit card then according to the law it's your credit card company's problem to solve. You phone them up an say you've been conned and they have to give you a refund. You can also moan to the local Trading Standards people. If you’ve been buying from Amazon you can write to them and complain, using the phrase "passing off" and threatening lawyers somewhere. That's legal-ese basically saying Amazon are at least partly culpable in a criminal offence of passing off a con as a genuine part. If enough people wrote letters like that then Amazon would soon do something about it.

Most of the problems arise from people not caring enough about the $5 they wasted on a fake part. Caveat Emptor. If people are too lazy to defend their rights then they'll lose them.

4
10
bazza
Silver badge

Re: Goodbye FTDI

This is exactly what I have done since the first incident - proactively avoided anything with an FTDI chip in it.

That is an ambiguous statement. Do you avoid FTDI's driver too and buy parts that have their own drivers? Or do you buy parts that continue to use the FTDI drivers?

The former is kinda hard to find, the latter is outright support for the knock off vendors getting a free ride on FTDI's back.

If you want something that works you're better off buying the former or buying genuine FTDI parts.

Personally I can't see what is wrong with FTDI's position. They can't be obliged to give driver support to clones.

Imagine the fuss if someone was making knock off Nvidia GPUs. If you purchased a $400 card that turned out to be fake you'd be moaning at the vendor until you got a refund. With FTDI no one can be arsed to get their lazy butt out of their chair and moan to their vendor about having been sold a $5 con. Weirdly they're quite happy to take a small financial risk and sponge off FTDI's living so long as it doesn't cost very much, yet moan like hell when things stop working and would have to buy a genuine replacement to get going again.

14
15

Oracle to kill off Java browser plugins with JDK 9

bazza
Silver badge

Sigh, that's a lot of Netgear's stuff out of date (some of their switches use the java plug in their Web interface).

I wonder if they'll fix it or simply launch a new range. I think I know the answer....

1
0

Safari iOS crashing: Suggestions snafu KOs the Apple masses

bazza
Silver badge

Re: I had it this morning

What's more interesting is that Safari is clearly not properly validating responses from Apple's suggestions servers.

That points to an opportunity for the man in the man in the middle attacker. If you can crash Safari with a dodgy suggestion, what else can it be made to do with some carefully chosen responses?

0
0

Apple growth flatlines ... Tim Cook thinks, hey, $80bn is still $80bn

bazza
Silver badge

Re: "growth in services revenue"

@mexflyboy,

you still think Blackberry has a chance in hell?

You stay in your data slurping walled garden with your shiny iOS/Android hand held penis extension, there's a good sheeple.

0
4
bazza
Silver badge

Re: "growth in services revenue"

Hmmm ... You most definitely have had a quite different experience than I've had with my iDevices: I haven't experienced any (that's "any," meaning "the slightest bit of, a scrap of, a shred of, a whit of, a particle of, an iota of, a jot of") the intrusions of which you speak.

Odd, eh?

I'm wondering if once you're inside Apple's walled garden does the nagging go quiet?

2
3
bazza
Silver badge

Re: "growth in services revenue"

Any good CEO's favourite motto: once you've got them by the short and curlies, squeeze, squeeze hard. Walled gardens should have a warning sign, "Abandon all hope ye who enter here".

BlackBerry

Try a BlackBerry (running BB10). You get none of that kind of shit with their music player, etc. And they have a neat cloudless way of syncing music and stuff between PCs and phones. The lack of intrusiveness will feel like a breath of fresh air. Plus you get their legendary messaging client, which is still the best out there.

4
5

'Unikernels will send us back to the DOS era' – DTrace guru Bryan Cantrill speaks out

bazza
Silver badge

Re: "Operating systems these days..."

There was an attempt to do so. It was called OS/2.

They should have waited for the 80386...

4
0

Boeing just about gives up on the 747

bazza
Silver badge

Re: Museums

Funny how it managed that when Concorde was never allowed to fly supersonic over the USA .

The SR71 flew a lot higher, and that does a lot to help dissipate the boom before it can hit the ground. They were very careful with their overland routing too. But even then they still could caused problems, and the USAF had to occasionally apologise, promise not to do it again, etc.

In contrast Concorde operating a daily service would have been a guaranteed problem every day - much worse than the occasional cock up by the military.

One of the SR71 pilots tells a story about cruising alongside a Concorde, and being waved at by passengers and being toasted with raised champagne glasses, at Mach 2. Surely one of the coolest formation flights ever! Both were astonishing aircraft.

2
0
bazza
Silver badge

Re: Twins are the most powerful.

It had an upside, though, takeoff's in 757s were impressive. It only needed 80% thrust normally, but if the pilot got permisison to use 100% for weather or similar reasons you really noticed that kick in the back :)

Ah, the majority of 757s ended up with Rolls Royce's engine. The thing is is that RR offered a derated RB211 - it was significantly over powered for the 757. Of course they didn't derate it down to the minimum power required for the aircraft, they made sure that it still had plenty of get-up-and-go.... Frank Borman, the famous astronaut and President of Eastern Airlines said that the 757's RB211s were the finest aero engines ever made.

The 757 in general was responsible for the development of ETOPS, and laid the foundation for the modern airliner industry. A very significant piece of work indeed.

4
0
bazza
Silver badge

Re: 747 favourite fact.

When one of those crashed after take off from Stansted they never found the DU. Must have burnt off in the fire. Nasty stuff...

0
0
bazza
Silver badge

Re: Twins are the most powerful.

Actually, twin engined aircraft are the most powerful of all. This is because the FAA requires a multi-engine aircraft taking off to become airborne and fly safely after the loss of one engine. Doing the simple math, with 100% used as the amount of thrust required to achieve liftoff and fly safely (use any percentage you want), it works out as follows:

The consequences for a twin is that most of the time it is flying two massively overpowered engines at a fraction of their max output during the cruise. That's not necessarily as fuel efficient as 4 smaller engines working "harder" for more of the time.

Leeham News did a good analysis of the very complicated trade offs you have to do to choose. One of the big factors traditionally was less maintenance man power hours for twins, but nowadays it's not so clear cut. Those enormous engines take a lot of looking after.

5
0
bazza
Silver badge

With Boeing not designing another 4 engined VLA, the eventual replacement for the upcoming 747-8's they're buying right now could be very expensive, or European (if hell freezes over).

With so few commercial operators of 747-8's, keeping the president's new aircraft flying for 20+ years could be very expensive... in 10 years time they will possibly be the only examples still flying.

0
0
bazza
Silver badge

Boeing don't seem to get cabin noise.

Well, they sort of do. The 787 can be quiet if the airline buys the optional and heavy sound proofing. A lot of them don't fit it in economy so the 787 ends up being so-so. It doesn't help them when they squeeze 9 seats into each row...

AFAIK Airbus don't give the airlines the option, sound proofing is standard, or is easier for the airlines to accept. They've also been better at optimising fuselage dimensions vs seats per row (A350 is 8 inches wide than the 787). You get the bit more space and a quieter ride.

With A380 (and A350) operators setting benchmarks for "comfort" a significant number of passengers are making purchasing choices accordingly.

For those airlines, eventual replacement of the A380 might be difficult. The new 777x isn't looking so spacious, and the walls are thinner (= noisier?). Both A350 and 777x can't carry anywhere near as many passengers. Boeing aren't doing another VLA. If no one orders A380s it will go out of production.

They won't want to get a noisier or more cramped replacement... Buying more A380s whilst they can might be a major factor in how orders are placed in the near future.

4
0

Zuck slapped down for privacy breaches in Germany again

bazza
Silver badge

Now What?

It means that any similar new start up can't pull the same trick. Facebook benefit from having done it earlier.

How is that going to be corrected?

And what about Android looking up callers' caller ID? It's a similar third party data slurp without consent that gives zero benefit to anyone but Google. That'll have to go too.

2
1

Eight-billion-dollar Irish tax bill looms over Apple

bazza
Silver badge

Re: Actually, what about the loans?

Apple borrowed a ton of cash in the US to pay their shareholders a dividend. Their foreign cash holding was used as collateral. Now that collateral might have to shrink by 8 billion those loans might have to be reconsidered.

Not that it will matter to Apple. But it might to other US companies that have used the same trick and have cut things closer. They might get a sudden and unwelcome bill...

10
1

Aircraft now so automated pilots have forgotten how to fly

bazza
Silver badge

Re: The human pilots just do the easy bits...

I think you'll find that most of the pioneers from the 1960s and 70s are already retired. A lot of them are dead and buried.

Things are very thin in IT. How many people can design silicon chips? Hardly anyone. Even Intel struggle. A lot of their designs are Israeli, because that's where the guys who can do it live.

7
6
bazza
Silver badge

Well, Duh...

There has been concern over this brewing for years. Hang around pilot forums for a while and you'll be amazed it's taken this long for a report to be published.

The regulators have been fairly weak over this. There has been a lot of lobbying by airlines... The level of pilot skill back in the 1980s formed part of the safety case for this level of automation. By allowing the airlines to slacken off they have weakened that safety case. However they can point to air travel being safer overall now than it was before all the automation came in. Humans make a lot of mistakes, and there has always been a lot of diversity in pilot skill levels. Automation has prevented a lot of cock ups.

The trouble is that a few recent accidents have happened to basically airworthy planes with minor or no technical defects.

The problem varies between airlines. There are some airlines out there who put pilots on disciplinary measures if they turn off the autopilot between taxiing onto and off the runways. Guess how well their pilots can fly...

Also you could take this report, swap "pilot" for "driver", and you basically have the report that will be published 10 years after the introduction of self driving cars.

12
0

No escape: Microsoft injects 'Get Windows 10' nagware into biz PCs

bazza
Silver badge

Re: A Question of Consistency?

There's going to have to be a backlash at some point.

Online advertising in the UK costs every working person £240 per year (see here and here; £7.2billion / 30million = £240ish). They pay for that through the price of goods in the shops. The Consumer always pays.

The likes of Google, etc, depend on growing that. But at what point does it become unsustainable? At what point does the population start turning away from advertised products, throwing their devices in the bin wholesale? Will that happen when online advertising is worth £1000 / wage earner, £2000, £10,000? £10,000 would be like having a second mortgage for most people. Yet Google's entire reson d'etre is to achieve that.

I reckon if you asked someone was Google Maps worth as much to them as their house, they'd say no. Half? Tenth?

At some point the sense of proportion will have to kick in, especially if the economy goes wrong again.

1
1

All Systems z are Go: IBM ports Google language to mainframes

bazza
Silver badge

Re: Singing the BlueZ?

"Except unfortunately IBM."

Depends. IBM started putting a lot of effort into Linux specifically so that they could sell more mainframes.

According to the article, they've sold more mainframes. Mission accomplished. Though there may well have been some unintended consequences...

0
0

Trustworthy x86 laptops? There is a way, says system-level security ace

bazza
Silver badge

Re: Trusted storage

Has everyone forgotten that some of Sun / Oracle's SPARC CPU designs are open source?

If you really wanted to you could inspect and approve their design, set up your own fab, make your own chips, build your own machine, and put Linux on it. If your not interested in the smallest possible transistors (14nm today?) the fab could be comparatively cheap.

You can have a complete open source hardware software stack without having to start again from scratch.

5
0

Chat messages in Skype for Windows are bang out of order – so here's how to 'fix' it for now

bazza
Silver badge

How?

What on earth have they done in their source code? To take existing code, which presumably does nothing more than deliver messages in the order they were sent, and change it so that instead it stores them up and displays them out of order sounds like extra work.

Surely at some point someone should have thought, "Why is there more code than there used to be?".

6
1

Getting metal hunks into orbit used to cost a bomb. Then SpaceX's Falcon 9 landed

bazza
Silver badge

Re: Unreservedly, unabashedly euphorically optimistic

They will figure it out. They'll make the unavoidably expendable pieces inexpensive and quick to replace.

The bits that are most prone to trouble are the turbo pumps. These are effectively the only moving parts in the whole propulsion chain, and they get a hell of a hammering. It took NASA a lot of effort to get those right on, well, every engine they've ever had, especially the Shuttle. They're far too expensive to be expendable, and are also a key part of the innards of the engine.

I don't know why people are getting so worked up about the knowledge now 'available' to SpaceX following this flight. You don't need to fly an engine to see what it looks like after a burn, you can pretty much do that all on the ground on a static test. They probably already know roughly what maintenance work is required to re-use the engines. All SpaceX need to do is see if the flown engines match static test engines.

And assuming that every engine is static fired anyway before it flies, every engine in a sense has already gone through the required maintenance regime to return them to flyable condition.

[As far as I'm aware the only engines in the history of space launches that weren't static fired before flight was some of the upper stage motors on Ariane 4, later on in the lifetime of that launcher. They had become so good at making them (they traced their origins back to Blue Streak) that there was little point in test firing them. Maybe the Russians don't bother any more either, given the age and success of their design.]

I suspect that SpaceX will very interested in how the structure of the 1st stage has coped with the battering it's undergone in making an about face turn, a slightly slowed ballistic return followed by some sharp deceleration just prior to landing. There's also the acoustic and heat loading it will experience on landing on a flat concrete landing pad. On a launch pad these are taken away by the flame trench (I don't know if SpaceX use sprayed water to suppress the acoustic load). On the plus side the amount of thrust needed to land is way less than that to take off - there's a lot less weight!

7
0
bazza
Silver badge

Re: A bit negative...

SRBs are basically a metal tube.

There's very little in them that's breakable - other than gaskets that were expected to be replaced every time.

They've also got a lot more power. Simple, cheap, and a huge amount of grunt - what's not to like about them?

They do have one major disadvantage - they're not smooth burning. If you've got a lot of SRBs attached to your launcher you get a rough ride whilst they're burning, and it gets rougher as they burn (there's less weight to be shaken about). That's not necessarily a problem, you simply tell your customers how strong their satellites have to be. However, it is an unwelcome complication.

2
0

SpaceX launch is a go for Sunday after successful static fire completed

bazza
Silver badge

Re: Management vs. engineers? It's not black and white.

The recent news from SpaceX shows an outfit which is run according to the principles of good engineering management.

I have to disagree with you. As you say their last failure was down to poor quality control practices which lead to a below-spec strut being launched, according to their own publications. And prior to this launch they're finding problems with a very cold throttle valve.

In this day an age in no well run engineering environment launching rockets (and aspiring to launch people) should the strength of a vital strut be unknown, or a vital throttle valve be found to be sticking that late into the launch process. It's not like the temperature profile of a valve was going to be unknown before they filled the thing up with cryogenic fluids.

Fundamentally speaking SpaceX are not carving a new pathway for how these things are done. NASA and the Russians did that, and Ariane too to some extent, working out that total quality control really, really matters. On a machine where every part is essential, it is essential that every part is spec'ed, built and tested to ensure that it works as needed.

To me it feels like SpaceX have found that there are not huge improvements to be made in design (they kinda tried that, turns out a lot of rockets are fairly efficient thrust / weight) or manufacturing (they tried that, turns out you cannot make the manufacturing process a lot cheaper). If you can cut QC processes, that's a big saving (and may be they've been trying that). If you can re-use large parts of your rocket, that too is a big saving. Their recent success is a big step towards potential cost reductions. They can consolidate that advantage if they apply QC sufficient to make it routine, not lucky.

0
0
bazza
Silver badge

Well, they pulled it off, congratulations to them. Now, can they do it several times in a row?

"A sticky throttle valve being dealt with on the pad by fiddling with the launch sequence?"

Welcome to the real world. Don't tell me this kind of think isn't done in state space programs.

Er, no. To succeed in the real world you write a specification, and build and test against it. If you're discovering hitherto unknown issues that late on then something (the spec, the build or the testing) earlier in the process has gone wrong, and you're launching equipment whose performance is, by definition, unknown. You are then trusting to luck, not an option in manned space flight.

"A wise astronaut wouldn't think twice, they'd drive home.

No-one drove away from NASA though.

No they didn't. And before the Apollo pad fire tragedy everyone including the astronauts knew it was appallingly risky.

It's still hideously easy to die during spaceflight. No one can afford to take short cuts or get complacent. Just because we've got 21st century kit doesn't mean that anyone can make any assumptions about how well it will perform. We still have to get specifications, build and test completely right.

0
0
bazza
Silver badge

SpaceX is now fiddling with the ignition timing sequence to see if that will fix the issue.

I'm sure SpaceX wouldn't say it quite like that... Nevertheless it does sound a teeny weeny bit like last minute discovery of how some of their equipment actually functions, again.

I have to say that optimising the timing of events during the launch preparation and ignition does sound a little bit like they know the throttle valve is not good enough. I wonder if the payload insurance company is paying attention; premiums will be going up, even if the flight is successful.

Ready for Manned Flight?

It also doesn't bode well for manned flight on SpaceX. It's one thing to fiddle with things like this when there is little but a well insured customer payload to lose. Any sober minded astronauts paying attention to the condition of their launch vehicle would want to take a long look at that, and all the other "experiments" being tried at the same time.

First rule of manned spaceflight; first you must persuade some of the best educated, most competent and innately sceptical people on the planet that the launch vehicle that you've built is going to work properly. Fail in that persuasion and you haven't got manned flight; they won't get into it. If you set out to hoodwink them, well you'll likely be having to explain that to a board of inquiry, a bereaved family and in this day and age possibly a judge and jury.

Astronauts know that all of the West's losses in spaceflight have been ultimately attributed to managerial failings. It was known that a pure oxygen atmosphere was a dangerous idea. It was known that measuring the circularity of a booster segment at only six places wasn't good enough and that launching at such cold temperatures would mean that sealing rubber wouldn't be pliant. It was known that insulating foam got stripped off the external fuel tank and could hit the shuttle during the launch climb.

A sticky throttle valve being dealt with on the pad by fiddling with the launch sequence? A wise astronaut wouldn't think twice, they'd drive home.

The stakes are high; SpaceX can ill afford another failure

Hmm, one wonders when they'll learn that patience and thoroughness are ultimately cheaper than haste and carelessness? I wish them good luck of course, but ideally they shouldn't be needing it.

12
14

'Unauthorized code' that decrypts VPNs found in Juniper's ScreenOS

bazza
Silver badge

So That's How Well...

...their code review process works. Seven years.

Sigh...

For a company whose entire offering relies on them having a reputation for doing things properly, they don't seem to have done that very well. They've just learned that their system of code review has been pointless all along.

So what else is lurking in their code base, how are they going to find it, and what procedures are they going to impose on their staff to ensure that this doesn't happen again? Proper multi-team clean room review is very expensive, but not half as expensive as destroying your entire business by going without it and running into something like this.

10
0

Nokia, ARM, Enea craft new TCP/IP stack for the cloud

bazza
Silver badge

Re: TCP Offload?

Hmm, as the article relates there's plenty of problems in the land of video streaming that this kind of stack helps solve.

I've never had a chance to benchmark VxWorks' IP stack. Given how light a touch the whole OS has it stands a chance of be pretty quick. INTEGRITY runs a stack in user space too. In fact, apart from Linux (and Windows?), what OSes run the stack in the kernel?

If they are quick then it's an interesting balance to judge; expensive but fast embedded OS that saves us from writing code, or bare metal coding that involves a lot more effort but much cheaper / free licensing. Well done for getting max performance!

0
0
bazza
Silver badge

TCP Offload?

I'm wondering how a TCP offload NIC fits into such a scheme. Having user space code accessing the NIC for this purpose sounds fiddly.

And wouldn't it increase the attack surface of the NIC too? Or has TCP offload gone out of fashion anyway?

1
0

Ofcom retreats from 4G spectrum auction after legal threat from Three, O2

bazza
Silver badge

Re: Hmmmmm

Any self-respecting filter should be able to sort that out...

Well we'll see. I'm not convinced that every £5 WiFi NIC comes with a half decent front end filter.

...and it's not like wifi is below the noise threshold like GPS is,

Ah, a classic mistake. If an RF front end gets saturated by a high power jamming signal, the receiver won't be able to pick up any other signal at all regardless of whether or not it's above or below whatever one defines as being the noise floor. A front end amplifier that's clipping won't pass anything worth a damn.

So if a WiFi NIC has a poor quality RF front end (a distinct possibility) and is co-located with a 4G phone (effectively a certainty in this day and age) pushing out a good 0.5W of RF, it's by no means certain that the NIC won't get saturated and crap out.

I'm sure OFCOM has thought about it, or at least I hope so...

2
0
bazza
Silver badge

Hmmmmm

Two bands were made available (2.3GHz and 3.4GHz)

So how close to WiFi is the first one of those going to get? And how many crummy WiFi devices are there with awful RF front ends with next to no selectivity that are going to crap out?

Lightsquared all over again?

7
0

Beardy Branson bangs birds on Boeing

bazza
Silver badge

Re: Prelude to Space

Apparently the whole idea works a lot better if you fuel it with Americium instead of Plutonium or Uranium.

2
0

Taxi for NASA! SpaceX to fly astronauts to space station

bazza
Silver badge

Re: PR luvvie

"One successful launch or test does not prove the design as you may have been lucky that something operating very close to its limits didn't break/malfunction (imagine the number of variables that change between launches). A proven design requires more than a single successful test."

That's nonesense. If a launcher successfully makes it to orbit then it is successful. Everything else you're referring to is associated with poor quality control, poor manufacturing records, and a general inability to know what actually got built and launched. There's no room for luck in the space business, especially if you're going to ask someone to be launched into orbit on top of it.

Dealing with poor quality assurance processes by building in extra margin is not an option for something as weight concious as a launcher. At least it isn't if you want it to be able to carry a worthwhile payload into orbit.

0
2
bazza
Silver badge

Re: riding in one of the safest, most reliable spacecraft ever flown

I remember reading a long time ago that Arianne was deliberately built with manned flight in mind. Whether that was a ruse to get the engineers to do a proper quality job or not I don't know, but the end result is certainly very impressive.

There have been brief flirtations with ideas such as putting a mini-shuttle on top of an Arianne, but there's never been a strong enough financial / geopolitical reason to actually go ahead and do it. Russia's manned launcher is just too well established.

Even the Americans are not strongly motivated politically to get back into putting men on rockets. They're doing it very slowly and always with a beady politician's eye on the budget.

3
0
bazza
Silver badge

Re: PR luvvie

"But spaceX still have to do the inflight abort test, and prove that the strut issue is behind them by doing a few more launches."

Sure, they have to get some successful flights off the ground, but there's more to it than that. The last flight failed because their quality assurance wasn't up to scratch. SpaceX said as much. If they've done nothing to change that, and merely decided to take a closer look at that strut, then no one knows what other problems are lurking in their manufacturing build chain. If they had poor quality assurance on struts, what else were they not looking at properly?

Effectively the next few flights are a test of the processes they have for building a rocket, not really a test of the design itself. Once you've had a single successful launch (which they have), the design is proven. Every failure after that is down to the inability to control the build of that design, which is seemingly where SpaceX have been for some time now. And if they've still not got the right build quality assurance process then their rockets will continue to fall out of the sky.

Of course, SpaceX aren't stupid and I'm sure they know this as well as anyone else. I hope that they have got on top of their process problems. If they haven't, someone will eventually have to call a halt to their activities if they haven't gone bust first. You can't keep launching rockets over people's heads without a satisfactory approach to quality control, not in this day and age.

Russians

The Russians are still learning things about the way they build their 1960s design. The routing of a pipeline was thought to be non-critical until quite recently. Then they built one where this pipe, purely by chance whim of the construction engineers on the day, shared a bracket with another. The fact that one of them was carrying a cryogenically cold fluid and would freeze the fluid in the other (the heat being conducted by the bracket) was something they discovered only after it failed and crashed... Needless to say the pipeline routing has been rethought.

Arianne

The first Arianne 5 suffered a navigation failure due to dodgy re-use of Arianne 4 software (the spec was fine, but the actual implementation didn't quite meet the spec. Very embarassing). It was blown up by the range safey system. They've had only 2 out of 83 outright launch failures, and those were both a long time ago in 2002. They've been sending them up regular as clockwork for years now, and from what I hear they very rarely suffer even 1 second of launch delay (except for weather).

Commercially speaking Arianne is a rock-solid proposition; slightly more expensive than, say SpaceX, but your $billion satellite will get launched on time into the right orbit with the cheapest insurance premium in the business. That commercial certainty is worth a fortune to the companies that commission and operate satellites, far more than the actual cost of the launch (thought to be €100million for Ariane 5).

All the time your satellite is sat on the ground waiting for the launcher guys to get their act together you're losing a monumental fortune in lost revenue, interest payments and programme costs. A cheap launch that's 6 months late is hugely expensive for the satellite owners.

10
0
bazza
Silver badge

PR luvvie

"When Crew Dragon takes NASA astronauts to the space station in 2017, they will be riding in one of the safest, most reliable spacecraft ever flown."

Sounds like PR bullshit to me. The most reliable rockets ever flown are the Arianne 4, 5 and Soyuz launchers. There's no way Space X can launch enough flights in the next year to gain a statistically significant success rate to even begin to compare themselves with the others.

Richard Feynman had a thing or two to say about PR and space. Quote:

"For a successful technology, reality must take precedence over public relations, for nature cannot be fooled."

I hope NASA and SpaceX haven't forgotten that. And I hope that whoever volunteers to sit at the top of that thing doesn't forget it either and goes and takes a real close look at SpaceX's quality assurance procedures. As other people have pointed out SpaceX haven't got a good reputation for reliability yet. SpaceX have themselves admitted that their last failure was down to inadequate quality assurance. Have they really managed to fix all theirs problems of that sort in just 5 months? Hmmmmmm.

16
1

Microsoft quietly slips out patched patch for Outlook – in camouflage

bazza
Silver badge

Re: Unit testing

Unit tests are a great idea in principle, but how do developers emulate all the possible permutations of code on their development systems?

Who's code are you talking about?

MS's APIs have a published interface, and all that their unit tests have to do is test that their DLLs, etc. satisfy that published specification. They don't need to care what use applications make of that interface, just so long as adherence to the specification is enforced (parameter range checking, etc).

From the description of the bugs that were arising it seems that their patch changed the interface, or just didn't work at all, both of which would have been revealed by an adequate unit test. They could have replicated Outlook's and the login screen's use of that interface in the unit test if they'd wanted to be sure they didn't break their own applications, but it looks like they hadn't done that.

3
1
bazza
Silver badge

Re: Moving faster

No, the pace needs to slow down. Nobody can handle the current pace of development.

Seconded.

Strategy Anarchy

MS's strategy has been all over the place in recent years. That makes it very hard to know what the best way of developing for their platform is. Metro was a disaster for those few who took the plunge...

Alternative Strategy

A loooooong time ago MS showed Windows 7 + Office 2007 compiled and running fairly satisfactorily on an ARM dev board. It was even printing to an Epson printer. All they'd done was an ARM HAL for the NT kernel, recompiled the entire thing for ARM, and run it. Unsurprisingly it worked.

If they had followed through with that they could have had an ecosystem of fat x86/ARM binaries with an option for developers to put a touch interface on their applications too if they thought they'd be used on a phone or tablet as well as a desktop.

For the developers that would have been great - one source code base, a choice of supporting one or two application interfaces to suit, running on either desktop x86 or ARM mobile.

For users it would have been great. Your mobile would have a mobile interface when on the move. Plug in power, a HDMI, bluetooth mouse & keyboard and voila, a full desktop with desktop apps.

Why they didn't do it

The engineers who did that demo of Win7 + Office on ARM must surely have seen the potential for a strategy such as I've outlined above, and must surely have been desparate to do it.

The barrier at the time was that ARM SoCs weren't quite up to running a full fat desktop OS kernel on battery power in the memory (512 MByte was typical).

Those problems don't exist anymore, haven't existed for a few years now, and were always going to disappear as the ARM SoC arms race took off. ARM has had a good dose of Moores law...

Had MS chosen that strategy they'd be doing veeery well now. Instead the impatience to be seen to be doing something, anything, in the mobile space meant shoe-horning a nasty cut down Windows onto an ARM powered devices and porting some of that nastiness and cutdownedness to desktop (Windows 8 / Metro).

Where did that impatience come from? Ah yes, shareholder pressure... There must have been engineers weeping into their coffees.

I suppose one advantage that Apple had was that Jobs ignored his shareholders (wouldn't even pay them a dividend), whereas MS can't quite do that.

2
1
bazza
Silver badge

Re: Why after all these years....

Try to build a testing system that can handle all the variations in base hardware, devices, software add-ons. Try testing all those permutations against all the system software that has been on the x86 (or wherever) since 198x.

Try running the unit tests on the modified code and release it only if it passes. That's what a unit test is for, to make sure that a piece of code performs to specification.

Field issues like this mean either that their unit tests for this piece of code aren't adequate, or they didn't run them.

I like Windows generally, specifically 7, still thinking about 10. I have to say that over the past few months it does feel like MS have not had their A team working on patches, which is not exactly good news for me or them.

I've thought long and hard about Linux, but Libre Office is too slow and not very good, the distro anarchy is a massive put off, there's only one decent-ish desktop (Cinnamon, but that has bugs and annoying features), there's the whole systemd thing.

I also wonder where Apple will go with OS X. Will it still be there in 5 years time? And Apple aren't exactly brilliant at patches or framework stability either. Plus the hardware is stupidly expensive...

3
3

Big Bang left us with a perfect random number generator

bazza
Silver badge

Re: Passing of Random Data

One way of looking at a perfect file compressor is that, without knowing the decompression algorithm, its output should appear to be a stream of random data. So you can always claim that random data is not an encrypted message, it's merely a message.

The fact that it doesn't decompress in ZIP, etc. is neither here nor there.

Of course, and alas, ZIP and its equivalents cannot actually get that close to the Shannon limit.

2
0

Samsung S6 calls open to man-in-the-middle base station snooping

bazza
Silver badge

Oh Good Grief!

Please, will someone somewhere just implement something properly, just for once?

10
0

Facebook CTO: Clear legal grounds needed for EU-US data exports

bazza
Silver badge

Re: "Most Europeans are sharing their data with people outside Europe"

Where it gets trickier for Facebook and indeed Google is data they collect on people who are not users of their services.

I'm referring to Facebook's habbit of tagging people in photos even if they're not Facebook users. Thus they can build up a social network for such people even if they themselves never have any direct contact with Facebook.

I'm also referring to Google using caller ID on Android phones to build up data on who you are and which Android users you call, even if you yourself do not have an Android phone or use Google in any way. Google get to know who you are if you're in the callee's contacts list and can of course Google your phone number.

In collecting data this way Facebook and Google are really skirting on the edges of legality. They have no terms and conditions to point to to say you've signed up, and they're currently exporting such data abroad. Should anyone in Europe choose to sue over being surveilled in this manner, they'd lose almost instantly, especially in the current climate.

Good job for them that there isn't really an equivalent of the American class action system in UK. There is a Information Commissioner who, should they get angry enough, can bare some pretty sharp teeth.

1
0

AMD sued: Number of Bulldozer cores in its chips is a lie, allegedly

bazza
Silver badge

Re: A bit of a Dickey move

I remember that the early Sparc T CPUs from Oracle / Sun had eight cores that shared an FPU. Never heard of anyone suing over that.

The first Nahelem architecture Xeons were "deceptive" over many SSE registers they had. In 64 bit mode all the registers were there as billed, but the number mysteriously dropped by half in 32 bit mode. It made 64 bit mode code look terrific compared to 32 bit, but only because of this artificial limitation. Of course this was all written down in the data sheets...

10
4
bazza
Silver badge

Re: A bit of a Dickey move

As far as I know, Intel does not sell hyperthreading as more cores.

But they do drop very heavy hints that it makes things faster, which is not necessarily so.

It'd be crazy if this case is allowed to get to court, never mind win. If a sharing of resources means one cannot count all the cores sharing them, where does that end? Cache? Intel have shared L3 cache. Power? Memory? AMD actually do quite well there. PCIe bus? Not likely to be more than one of those. It's plain nuts.

SMP as we know it could be outlawed by a court case, which would be the most ridiculous thing ever. So it's probably guaranteed then.

16
5

Forums

Biting the hand that feeds IT © 1998–2017