Posts by Hugh McIntyre

Re: I understand

RE: "LE wanted the shortest validity time to reduce risks with temporary hijacks or expired/sold domains but compromised on 60 day renewals to reduce load, with a 30 day grace period to allow for temporary outages and other intermittent failures."

This is nice (and I use LE). But I can't help noticing that the Let's Encrypt Authority X3 signing certificate is valid from March 17 2016-2021, i.e. 5 years.

Not sure if this new 1 year limit from Apple will only apply to the leaf certificate or also all signing certificates up the chain? The latter is potentially more painful.

Exactly. I guess we will see what is actually proposed.

If the solution is that as long as you have at least one modern device then all of the legacy devices can just grab a slave copy of the audio from this one then you would still have whole-home audio, and the modern device can deal with any new audio services and OS updates.

If not there are going to be a lot of very unhappy ex-customers.

I actually doubt this was a cash grab or at least not thought out well if so. I think Sonos are a lot more worried about new potential customers buying Alexa and similar smart speakers. At least this seems to be why there are so many recent software updates which are all irrelevant to those of us who are almost always just streaming local music from a NAS.

Re: Why indeed...

Reading the list it's looks like all of my devices are probably legacy... at least it will be consistent.

In terms of audio quality there's a difference between the speakers versus ZonePlayers driving a regular amp and HiFi speakers. In the latter case the ZonePlayer essentially replaces a CD component and the sound seems to match the HiFi component. Meanwhile the speakers are not the same quality... In fact the only reason I have Sonos in the first place was to replace a CD player connected through a 1991 amp and Mission speakers; the sound is the same.

Presumably "no updates" will mostly be a no-op for those of us streaming music from a local SMB server, and only would be an issue over time if services like Pandora change their protocol and/or the local computer running SMB changes it's OS version and becomes incompatible.

Sonos used to have a "Bridge" device to convert from the local LAN to Sonos's private WiFi network, What they should have done in this case is to say that people may need to have one software-updated device on the network that can speak to new-protocol sources, and then this bridge can stream to other older devices on the local LAN. They claim there will be a scheme in May to separate old and new networks so we will see if they do something like this.

Certainly, there's no chance of me replacing all of my Sonos devices unless they come up with a transition plan, and I don't think I am alone.

Re: "I am struggling to understand what innovations have been made"

New versions are enormously faster. USB 1.x went up to 12 Mb/sec. USB 3.x can do 5-20Gb/sec, i.e. about 1000x faster...

Re: Been a Flickr Pro since they were founded...probably won't renew

SmugMug still supports upload from Lightroom Classic. (Speaking as a SmugMug subscriber). The SmugMug galleries are also more flexible than Flickr IMHO which was my original reason for switching.

Can't justify paying for both, hopefully the original SmugMug will continue at least.

I'm actually slightly surprised they didn't try to migrate the Pro Flickr accounts to paid SmugMug and try to drop the free accounts. Maybe they will in future ...

Re: Ex-loyal customer here

Re: "do they seriously think business' wouldn't want to back up vmdk and other files?!"

They may want you to run a copy of Crashplan inside the VM, thereby paying an extra license because this would count as an extra system? Yes I agree this sucks :(

Re: Ridesharing?

Lyft and I think Uber do have options to book a shared ride where you go with strangers (e.g. they pick up/drop off another person on the way). But most people pick a solo booking. For example this article says 35% of Lyft rides were shared last year: https://mashable.com/article/waze-carpool-app-ride-share-expands/

That particular article also says Waze has started real ride sharing service whereby a commuter can share with a stranger while still driving themselves, in exchange for cash for petrol/gas and presumably being able to use a carpool lane.

The bigger problem with Lyft and Uber is they need to make a profit at some point ...

Re: "GPL is cancer"

... and yet almost all of the commercial IC CAD software runs on Linux. Apparently they don't have issues, key libraries are under LGPL, and the fonts are fine.

Probably though this is because these packages previously ran con commercial Unix so Linux is the mainstream successor. Meanwhile other commercial software that didn't previously run on commercial Unix (e.g. Adobe software) don't run on Linux either.

So this is more of a Linux/Unix-family versus non-Linux/Unix split. Not GPL.

Even software that needs to include Kernel drivers such as VMware seems to manage to ship a commercial package without GPL issues.

Re: Not great.......

Intel bought Altera (who were using their foundry services), not Xilinx (who were not).

Re: Actually, this decision doesn't matter anyway.

A plaintiff could perhaps ask the court to get Twitter to unblock the blocked people?

More normally, other presidents (and even this one) have tended to comply with court rulings and/or appeal through the normal court system. But even if POTUS is inclined to straight-up refuse in this case, the request could go straight to Twitter. Imposing a fine, collectable by court order, is another thing the court can do without congress, although unlikely in this case.

Clearly this is not a "high crime" so provided the people get unblocked & no new people in future, this will get forgotten quicky.

Re: A lot of sites still sport self-signed certificates

Letsencrypt is free and not self-signed. No need for self-signed personal site certificates any more.

It does seem that the problem here is insufficient enforcement of SSL/HTTPS, unless the attackers were able to get fake SSL certificates by using a non-standard CA? The whole point of SSL certificates is that you do not trust DNS because the certificate says "website.com is 111.222.33.44, public key XYZ, signed CA_name". At which point if you trust the CA you should not be using a different IP address from fake DNS.

DNSSEC would be a good idea though, probably.

Re: No changes apart from the changes of course...

Re: "Uh, that does not make any sense. Why discontinue the only part of Flickr that produces some income?"

Probably not right now, especially for anyone who has both a Flickr Pro and SmugMug account who they will be happy to continue to bill for both.

As mentioned though, even the cheapest SmugMug plan is $48/year compared to $25/year for Flickr Pro. Do not be surprised if they try to migrate the Flickr Pro subscribers to the higher SmugMug price.

Re: I would agree with only LEAs having full access

There's also the usage (of whois) for individuals of, for example, "is this acme-service.com website associated with the real company, or some impostor?" But in that latter case you can also look at the HTTPS certificate if the site uses HTTPS and if they filled in name/address info in the certificate.

As for the rest, the opt-in part of Nominet's plan is reasonable (some of the rest may be debatable). Most registrars already offer a "hide registrant info" which personal registrations can use, so big companies that don't use this option are already effectively opting in to sharing, and hopefully other individuals defaulted to hidden. As such, responding to GDPR by saying "all WHOIS registrations move to hide-registrant mode unless people/companies affirmatively agree to non-hidden" seems like an easy choice even though the number of non-hidden whois entries may end up pretty small.

On the other hand whether paid-access-for-others stands up might depend on whether the domain owners opt in?

Re: Amazon? Postal Service?

Lots of times USPS gets used for delivery here. There's no specific way to choose one delivery service or not, except indirectly by changing delivery time.

Presumably Amazon picks whichever bids the cheapest price between UPS versus FedEx versus USPS versus their own delivery service. But also sometimes UPS packages also end up going into USPS for final delivery - apparently it's sometimes cheaper to do this than to send a UPS truck round.

But if you're ordering items that need signature this may restrict you to not-USPS. Or this may depend on where you live -- some locations may always be cheaper via UPS?

Re: "only way to transfer your contacts from an old iPhone to a new one is via the cloud"

Local password protected backup -> restore new iPhone from local backup has worked every time for me. No need for cloud, and all the same contacts are present with no need for a cloud backup for this case at least.

Now iCloud is useful to keep multiple devices in sync, but even then there's no reason why this cannot be just a per-user backup rather than being data-mined and combined with other users' contacts, which I suspect was Tim Cook's comment.

Re: There's a worrying implication

Last time I applied for a mortgage here the bank required permission to get the IRS to send them a copy of recent tax returns (note: not me giving them a PDF or printout of the return). If this bank had done the same they would presumably have found out exactly whether the income matched, not just relied on a PDF or printout that the applicant might have edited.

Re: Not going to happen

RE: "Chinese companies have been openly infringing copywrite products for decades. No companies have openly won any lawsuit against them. So why start now?"

I suspect this would be tariff barriers, not copyright (or patent) lawsuits.

To be honest several other countries have started with memory chips when building up local semiconductor fabs. Including Japan in the 80s as mentioned above, South Korea, and others. Even Britain's Inmos started with DRAM and SRAM as easier to get working first before logic. So working on local DRAM/NAND capacity is not a surprise.

Whether China succeeds in hurting other suppliers depends on whether their DRAM and NAND is cheaper than the existing companies which depends on who can keep in the lead for the newest technologies.

AT&T is already in the high speed Internet space.

The complaint is that if they gain ownership of HBO, Warner Brothers (movies/TV), and the Turner channels then they can then hold other TV competitors to ransom by demanding higher channel fees that get passed on to consumers. Also (in the complaint) that they could make it more difficult for TV-over-internet services like Sling to compete in future.

The antitrust complaint actually has some merit, since allowing content companies to merge with traditional TV/cable providers just at the point that cord-cutting alternatives are becoming more common (thereby kneecapping the cord-cutting companies) seems like a bad bet for consumers.

Re: Poor build quality to blame ?

Agreed.

Charge HR needed replacement because of bubble unglued from strap after 6-9 months. Then a second one failed the same way and got replaced by Charge 2, and it failed after ~ 3 months because of the charger dying.

I want to like FitBit, but if these quality issues are common it can't be helping the losses :(

Re: Gun, meet foot.

Best not to use "xhost +".

"ssh -X special@localhost" means you get the X11 display without xhost insecurity.

Re: Plus

You can also get a few notebooks with small screens that still have high resolution displays, including the 12.5" one I have here with a 1920*1080 display. The only problem is you probably also need glasses to see the small text :(

The batteries that ran out were for the locator beacon.

If the actual data storage for the recorders is in Flash memory then it should last a few years, at least assuming no damage to the IC packages letting in water or from mechanical stress. The AF447 recorder was recovered after nearly 2 years, for example, and cold water would tend to slow down leakage of data from the flash cells.

I do agree about the remote chance of finding it though. Someone may stumble onto the wreckage later, but as you say it's also possible it will be covered by a layer of silt and therefore eventually invisible. And it's suspected that the voice recorder wouldn't tell us the original cause anyway since it would not include the start of the flight when the unexplained maneuvers started. Similarly the data recorder may also just include running out of fuel at the end followed by descent :(

Hence the decision not to spend another $100M on an uncertain search seems understandable.

Re: I'm its a coincidence the plan includes a huge tax break for Trump

Kansas tried the same tax cut for "owner-operated businesses" 5 years ago, as well as other tax cuts, and it was an epic failure (nationally reported) that needed the taxes to be put back up this summer to pay the bills.

None of the claimed better growth materialized (performance was actually worse, if anything), except that many rich people avoided tax because of the same "small business" giveaway mentioned by Doug.

Hopefully the non-rich people whose taxes would rise with this week's republican proposal will sink this plan.

Re: Not The Real. Problem

Short answer: the question is ex-felons, not felons.

There are 9 states (including Virginia mentioned here) where felons don't get their voting rights back even after they are released and finish probation. So these ones would vote normally if they were not forbidden. 3 of the 9 states have small print automatically restoring first-time offenders or "minor offenses", but the general restriction applies.

Most of the rest do restore voting rights after release (15) or after parole (28). There are only 2 that apparently allow votes in prison (Vermont and Maine).

Re: Guests already do execute code on the host

RE: "Yes, but how often does that happen? Usually VMs are used as an easy way to manage multiple large applications or user enviroments on Windows platforms since Windows itself isn't very good at it."

There's also this thing called cloud compute where people want to run VMs securely, no ...?

Re: Just change the current tax laws.

Re: "If you sell advertising/software/etc. in France, that revenue accrues to the French subsidiary, and can't be funneled to <somewhere else>"

I think you meant "that profit accrues" but that's the problem - you sell some software for 100 Euros in France and the French subsidiary internally pays it's Irish subsidiary 99.99 Euros because the company says the software IP is "owned" by the Irish subsidiary. Hence only 0.01 profit in France and low French taxes on this 0.01.

Fixing this in general requires honest intra-company pricing which is hard to enforce, although countries could prosecute some cases to encourage honesty.

On the other hand if you really meant "that revenue accrues to the French subsidiary" then this is what happens today, so companies can choose which country shows the profit (same as today) or this becomes the turnover tax.

Possibly the right answer is percentage profit tax, i.e. if 10% of a company's revenue is in France then they would page French tax on 10% of their global profit regardless of inter-company accounting. This may be difficult though assuming different countries have different rules on what counts as taxable profit.

On the other hand if there's really no R&D in France then there's less added value and presumably less tax justified.

Re: Click Fraud

RE: "Which shows the fundamental flaw in pay per view and pay per click advertising. This type of fraud will continue until the gullible morons who place adverts stop placing ads on that basis."

Not disagreeing there's 'a problem, but unless advertisers buy ads on the basis of "please place adverts on theregister.co.uk, newegg.com, <other specific sites>" then they want some way to charge more when more copies of the advert are displayed. Pay-per-play schemes on Spotify or Youtube have the same risk.

Periodically advertisers have complained to Google about click fraud and demanded that "Something Should Be Done". So there is some effort to crack down, although right now this seems to be just treated as a containable cost of doing business. In particular the fact that the fraud uses a botnet is because it would be a bit obvious if all the fake requests came from the same IP address.

Re: H1B visas

Coincidentally the Mercury News had an article on this on Friday, at http://www.siliconbeat.com/2017/08/02/apple-h-1b-workers-average-139000-pay-outsourcers-dominate-visa-program-pay-far-less/

This claims 59,184 visas for major outsourcers (Cognizant, Infosys, Tata, Accenture, and Wipro) versus only 7,248 for Amazon,Google, and Apple for example, which seems kind-of lobsided. And it claims those 3 tech firms paid an average of $115K-$139K versus $72K-84K for the outsourcers.

So maybe treat outsourcers differently from H1B's for full-time engineering jobs? Although it may be difficult to define criteria for this :(

The other comment is that it's possible for both things to be true: regular employers struggle to hire junior/mid-level engineers (I have seen this as well as ckm5) but at the same time senior/older people can't find senior jobs, and/or there may be vacancies in software/RTL design but engineers with different types of hardware of software experience maybe can't find jobs.

Ideally there wold be a way to separate the visas for full time engineering jobs from outsourcers though, since the latter seem to be more of the problems.

Re: "must not have any known security vulnerabilities, must have the ability to be patched"

Re: "So..... if no known security vulnerabilities, why, errr, patch?"

Presumably, no known vulnerabilities when you buy the thing and needs to be patched if/when new bugs are found later.

Since the lack of patchability is one of the main problems of IoT, mandating the ability to patch seems like a good thing?

Re: KeePass to LastPass to 1Password

RE: "But at last check 1Password 6 doesn't support local vaults (forcing me to stay on 1Password 4)."

Nope, I have iPassword 6.7 here and have always used local vaults.

The only (main) missing feature with local vaults is that the sync only seems to work to mobile devices, not to other local computers :(. Very unhelpful.

I have to agree with the complaints about dropping the local vault version -- the fact that legacy 1Password keeps your data entirely on locally controlled systems is a major benefit. I would have recommended 1Password to others except that it seems impossible to get the one-off purchase any more, and who wants to recommend people sign up for yet another subscription?

Re: Satellites would do just as good a job against surface ships

I should have clarified:

Yes, I'm sure the Russians and others will use satellites in part to find surface ships. For one thing, oceans are big and you need to know where to start since sonar has limited range.

But satellite may not work in all cases -- clouds may mean you only see heat signatures which won't be as precise, you may not know which ship you are looking at unless you can track from source port, etc. So they no doubt *also* want the sonar signature.

Having said that, I'm sure they would get the signature eventually, so this is only a question of timing.

Re: memory encryption

Re: "Building a transparent hardware encryption of memory is conceivable, but I don't know of anyone who has done it."

AMD EPYC (Zen core) CPUs have hardware memory encryption, so someone has done it:

"Secure Memory Encryption (SME) encrypts system memory. Secure Encrypted Virtualization (SEV) isolates the hypervisor and guest VMs to prevent access to data in shared guest data areas."

More details are under http://www.amd.com/system/files/2017-06/Trusting-in-the-CPU.pdf. Some OS/hypervisor enablement is required, but no change needed for application software.

Re: "Barclays still make you have a debit card in order to use PIN-Sentry for online banking"

Not true, I have a pale blue card with "Authentication" on the top right which only works for PIN-Sentry, not ATM or Debit. It may be that you can't use a non-debit ATM card, but you definitely could get an Authentication-only card in the past at least. Contact your Barclay's branch ...

One of the complaints in the linked articles is that John Deere does not provide access to troubleshooting diagnostics. ("The EDL is the required interface which allows the Service Advisor laptop to actually communicate with the tractor controllers").

Perhaps what's needed is the same rule as cars, since all cars are required to come with OBD-II access and at least most of the trouble codes can be read with consumer-accessible OBD tools? Seems easy to extend the same OBD rule to other vehicles such as tractors?

In terms of John Deere they probably don't want their competitors using the same software in competing products. For example if John Deere spends a lot of cash developing some traction control software to make a tractor work better, they don't want a cheapskate competitor copying this. Even being able to see the software without copying may let competitors know what to do. Much of this was also possible in the past with mechanical reverse engineering, but probably not as easily.

It does seem though that updated sale-of-goods laws or OBD regulations may be needed to make sure people can reasonably diagnose products they have bought, either themselves or a requirement to provide service access to reasonably 3rd party repair shops at commercially reasonable prices, subject to not just copying the full software.

Re: Missing the obvious solution?

My first thought was also "why not just use a caching resolver, if the primary is not available?". If the default TTL is 1 day, small outages for common domains should be survivable.

But this proposal seems to be a solving a different non-outage problem i.e. ignoring malicious changes for common domains if you think it's more likely the domain owner didn't change it's IP addresses. For example people redirecting nytimes.com to a malicious IP address.

The problem remains how to distinguish intentional changes by domain owners from malicious ones. It does seem that signing DNS replies by the owner (with DNSSEC) would be a cleaner solution.

Re: Dear gods...

Most or all of the Integrated Circuit CAD packages run on Linux (only or at least mostly). It's true that package and board level design may run on Windows but IC tools use Linux. This includes Mentor's LVS/DRC and similar tools which are definitely up to date on Linux.

Re: 32/128Gb only

The iPhone 6s did the same - 32GB/128GB only, And the iPhone 7 only supports 32/128/256GB.

I assume the reason is to push people to 128GB since if people decide 32GB is too small and want 64GB, they will be forced up to 128GB since there's no way to add memory later. If they offered a 64GB version with a price in the middle, fewer people would pay for 128GB.

The 32GB version is definitely limited nowadays but Apple probably wants a lower spec for the entry-level model to keep the cost down, and the fact that 64GB is a better choice for many people will force them to 128GB :(

Re: Wait! What? They have our MAC Addresses?

Nope, don't see the local LAN IP address here from https://www.whatismybrowser.com/. Only the public DHCP IP address on the ISP WAN.

Some of the other info it displays clearly depends on JavaScript, so there may be more info visible with Javascript. Stack Overflow also claims that ActiveX on IE may give the MAC address (not an issue for those not using IE). TBH I'm slightly impressed the warrant knew to ask for the MAC address, but this is probably boilerplate request language from other computer warrants. For example, Google does not have payment info for most of us, I hope, but boilerplate ISP warrant language might ask ISPs for MAC addresses.

@ David Nash

Re: "How on earth did an online backup service (if that's what it was) allow content to be indexed by Google and accessible to anyone without credentials?"

I assume from the article that the files were backed up to something like OneDrive/GoogleDocs/Dropbox, i.e. not a real on-line backup service.