Microsoft lobs Windows 10, Server Oct 2018 update at world (minus file-nuking 'feature') after actually doing some testing Is the Windows October 2018 update here again? Did it ever exist previously? Are we all in a feverish dream where the latest version, build 1809, is stable and fit for purpose, and Patch Tuesday was totally uneventful? Our finger hovers over the "no" button, but we live in hope of someone one day fitted a "yes" key. Huawei …
My favorite Sammy J. quote.
(Yes, I know he's a badass with lots of good quotes from so many badass roles, and he rightfully earns his dough. I'm not hating on the guy, and I'm not even trying to play hipster / "before he was cool"; I just like that one line best.)
Noticed extremely sluggish performance after upgrading 1803 to 1809 last night, then checked the Intel processor speeds.
Installed on an older Core 2 Duo Processor (test machine) using MS update tool and the processor speed upper liimit was capped to 1.1GHz down from 2.5GHz. A quick reset of the power settings in the old Control Panel seemed to do the trick though but how many people will take this as their computer is just too slow to run Windows 10 and buy a new one?
Am I sensing a deliberate Apple Style processor throttling, to boost sales of new Win10 devices?
(I suppose we should be thankful it can run on a processor from 8 years ago, unlike Apple's macOS Mojave that is hard blocked from being installed on older <2012 models).
Installing macOS with all it's caveats (in terms of older hw), is actually harder than installing Windows 10 today, never thought I'd say that.
The snide comments in this 'article' would have some merit if you actually described your testing credentials. If you had an inkling of what it takes to test something this complex you might take a less sarcastic approach.
I was Defect Manager on a 3 year project of considerably less complexity and had to 'handle' >25,000 defects. Trying to understand the impact of some vaguely written report is not a trivial task so I feel for what the Windows team has to put up with.
"If you had an inkling of what it takes to test something this complex you might take a less sarcastic approach."
Or not.
Proper testing is a complicated and difficult task, there's no doubt. Which is why Microsoft decided to cut back on doing it. Given Microsoft's track record of software quality since the release of Windows 10, I think that a sarcastic response is letting them off easy.
"Proper testing is a complicated and difficult task, there's no doubt. Which is why Microsoft decided to cut back on doing it. Given Microsoft's track record of software quality since the release of Windows 10, I think that a sarcastic response is letting them off easy"
^ Indeed, and Microsoft really ought to consider unsacking all those quality control staff that they laid off four years ago.
@gerritv:
I've done many projects involving software as complex as Windows 10, easily, except defects weren't allowed to exist that might impact functionality, especially safety. That's why what I do is called safety-critical. The difference between myself and Microsoft is that if I fucked up, people were going to end up dead with millions of dollars of damage occurring. Billions of dollars wasn't out of the realm of possibility. The US Government might be a little ticked to come up missing a destroyer or Naval Air Station. I had to literally write bulletproof code.
The techniques involved are tedious in the extreme, formal verification just as an example, there are others involved. What consumed most of my time was not writing defect free software, it was insuring that the other code, the hardware, and especially the operating system itself couldn't be allowed to stab me in the back. I used to drop long lists of defects off at the trade shows to the various vendors asking them to fix their software.
So, I've zero patience for Microsoft's engineering processes, such as they are, when this kind of crap, that was reported to Microsoft by user-testing yet was released into the wild. Last point: I used to test their server software in alpha and beta status. For years. I can probably dig up my Beta ID if you'd like.
That really depended on a real engineer being around rather than just marketing drones. They appreciated the thought and I'd see their bugs slowly being removed along the way. Which is why I documented that damn things in detail so I could simplify the code. Then again, I'd pick up on others. Round and round it goes....
@Jack of Shadows - There are many industries were bulletproof code is the minimum requirement; programming errors could cause a cascade of other errors that could lead to at best nasty financial problems or at worst someone dieing. I work in the medical industry so I am attuned to the fact that mishandling data could lead to a wrong decision that could potentially kill the patient; generally a rare event but one to be avoided. But other industries face the same requirements of avoiding software errors that could cause deaths because the user relied on bad information generated by the software.
So like you I have very little patience with the farce of Bloat10 and Slurp's refusal to get head out of..
@gerritv
None of that is relevant.
They are a commercial operation who are paid to provide their products. No matter how hard it is, that is their job.
They have been making commercial decisions to save money, even though they made ~$16.5billion profit their last financial year, by cutting their test teams. They choose where to allocate their resources, and they could choose to allocate more resources to their testing and validation process. They could also choose to change their release cycle to yearly to more thoroughly test and validate their work.
If it's too hard, they are free to choose to exit that market entirely.
@gerritv - Slurp is solely responsible for proper testing of their products. When they obviously fail to do so they deserve to be hammered mercilessly until they get their act together. In reality, the harsh statements on El Reg are doing Slurp a service if they bother to listen. The natives are restless and do not need more reasons to bolt; they need reassurance that staying with Slurp is the correct course.
The monopolist John D. Rockefeller noted to maintain a near monopoly you have to give the customers good reasons to stay. Part of that is prices but also providing quality products. Low priced, quality products keep customers happy. Treat them with disdain either by raising prices or giving poor quality and you create an opportunity for a competent competitor to permanently take customers away from you. Rockefeller had legal problems with the ferals because of his market dominance not with customers.
MS deserve criticism for two reasons here:
1) Killing off internal QA that probably would have caught this bug prior to release.
2) The bug *was* reported in the feedback hub, and MS missed it. Hence, the feedback hub is clearly not fit for purpose and needs revising (something MS have admitted).
When you remove proper QA, then implement a "crowdsourcing" alternative that allows critical bugs to slip through despite being reported, you deserve some criticism. Especially when you "force" the buggy update onto people's computers because you've massively restricted users' control over updates.
Updating software to guard against security issues is very important. This is why it's important that users have trust and faith in the quality of updates. Regularly releasing buggy code and forcing it onto your customers machines is a great way to destroy trust in software updates in general, and that's dangerous territory if users start to increasingly see updates as something to fear...
This post has been deleted by its author
When you remove proper QA, then implement a "crowdsourcing" alternative that allows critical bugs to slip through despite being reported, you deserve some criticism.
When you do this and force the software onto people's machines you deserve to go out of business. I think Microsoft are displaying a level of conceited arrogance way beyond what their position in the industry will sustain. They are no longer the only game in town and do not possess whatever strength they perceive allows them to shit on their customers with rubbish like this.
It's exactly this kind of reason as to why I was so resistant to the Windows 10 concept in the first place. Had no choice but to end up using it due to work, and Linux isn't really an option due to a lack of desire to fight with every Windows game I want to install
However it does concern me that there's little to know repecussion for Microsoft, they can't even lose sales because they're basically giving it away....
My faith might be a (tiny) touch higher if they had upped the build number. Still calling it version or build number 1809 suggests they haven't rebuilt (all IDEs I have used automatically up the build number whenever I do a rebuild), which suggests a "your holding it wrong" attitude to the reported, rather serious, bugs. Even from a marketing droid's perspective, upping the build number by some significant value (say 42), suggests some hard work has been going on behind the scenes, which would in some way support their "we treat any data loss seriously" public statement (which in all likelihood should be "Seriously? We treat data loss?").
It's also just the wrong number. These 4-digit numbers used to be two digits from the year, and then the two digit month, which brought us such logical numbers as 1407 (July 2014), 1511 (November 2015), etc. Then they release a version titled the April 2018 update, which actually came out in May, and number it 1803. Hey, 03 means March, guys. From original release date and name, this should be build 1810, and from new release date, it should be 1811. Definitely not 1809.
Dear El Reg:
Don't make us send the boys from the Ministry of Love over to visit.
Repeat after me: The current update is the only update MS issued this Fall.
Oceania is at war with Eastasia. Oceania has always been at war with Eastasia. Right?
-- Your Friend, "Big"
@gerritv: testing pro replying here too (& a former MS fanboy might I add). This is a company with an awful track record of quality, which produces a full testing software suite & advises testing practices to others & who fird so many of their testing team! Serves them tight & this should be a lesson to all companies & project managers wanting to skimp on testing - that plan to 'save money' NEVER works out well!
Because it really doesn't matter for this machine. Annoying restore, yes, but sometimes I'll take it. So far, nothing apparent. Still bitching at me about reduced CPU performance due to firmware, thank you so much Intel. I'll kick the tires on Hyper-V later, that was the major change this iteration.
So tomorrow I get to install Octobers update and Novembers update on the one Windows 10 laptop under my control. Or rather, I get to watch as they get installed by themselves, coz I don't have full control. I can't wait for the higher ups to declare the nominal user of that laptop isn't actually coming back, so I can rip out 10 and install something saner. Then at least I can stick the admin password into my password manager, instead of not knowing it coz no one bothered to remember it. I only know the users password coz someone labeled it onto the case. Considering that particular laptop once took most of a week to install an update, I don't expect it'll get much actual work done on it tomorrow.
Meh, it's a charity, I'm a volunteer for only two days per week, the guy they pay to look after things should be doing a better job. There's a desktop running Vista that he should have installed 7 onto sometime in the last month, I was planning on doing that tomorrow.
I said it was a charity. There's stuff that has been hidden in cupboards for many years, just waiting for me to start working there and find it. I've had to fix up an ancient laptop that had Windows XP on it, so they could retrieve some important documents from it's hard disk. There was also a laptop that had Vista on it, but now has Linux Mint on it, though the display has lost a bunch of pixels in the lower right corner, almost ready to just toss that one.
"I've had to fix up an ancient laptop that had Windows XP on it, so they could retrieve some important documents from it's hard disk."
It would be much easier and quicker to open up that laptop, get out the hard disk and attach it to a USB port as external hard disk on any working computer. Just make sure you transfer those important documents to the proper folder on the network.
They may have done just that. Or as said "is a charity", so depends on if your personal kindness budget covers the fuel back home, or the cost of getting one a the PCworld next door/Amazon next day order.
If it boots, you could theoretically network into it, use a recovery disk, take out the HDDs, options are limitless.
"It would be much easier and quicker to open up that laptop, get out the hard disk and attach it to a USB port as external hard disk on any working computer."
But not cheaper for a charity that runs on donations and grants. I did think of that, but didn't have any suitable hardware laying around for reading that ancient hard disk, other than the ancient laptop it was in.
Will it bork my system?
Are they really, really sure that all my data and applications (including config files) will be there after applying the update?
If the answers are not 10,000,000% yes then ... stay well clear for at least a month (or more)
Is it me or does MS seem to be more like a bunch of numpties rearranging the chairs on the deck of the Titanic as if sinks between the waves? Nothing they do will stop the good ship Windows from sinking.
We got marked down the other day for some of our Server 2016 instances, as they hadn't got the latest patch applied - fuckwits.
You can't win, can you. You either roll-out patches immediately, and risk being an unwitting beta-tester, but be compliant, or you wait, and test, and wait for Microsoft to fix it, and then get called out for being cautious.
Gaaaaaah!
Critical patches within one month, everything else within an "appropriate time frame" (e.g. 3 months) - isn't that what PCI's documents and training courses state ?
Of course, Microsoft had to make that so much more difficult by bundling up patches and making it harder to apply/test just the critical ones.
It is like Ubuntu release names, first two digits are the year and the second two are the month. Although as they withdrew it and then re-issued it should probably be 1811.
Calling it 1811 would show that it wasn't released when they said it would be and that would involve admitting a fault so it is still 1809
1809 is just the public-facing version number (obviously based on the last two digits of the year and the month it was released in). It's just a name plucked from the ether for something specific.
Build numbers have a different significance compared to version numbers, generally.
There are three version numbers. The 10.[randomdecimal] one that doesn't matter, the build number, and the version number. The version number is 1809, even though that's the wrong number (see my post above for why it's wrong if you care which you probably shouldn't). You have the build number, which at least goes in order. So when that number goes up, the build is newer. Build 17134 seems to be the final version of windows 10 version 1803. So in short, it's crazy.
Welcome to the 21st century - I suspect that the Windows sources have become so complex with patches and features added and fixed and re-patched since it's release, that any changes are simple rolling the dice as to what they will break. Does anyone actually understand the whole picture any longer or is it simple one team working against a dozen other feature update teams?
You can evaluate and validate your teams patches and features but when another team rolls out its updates, who checks to see if there are any interactions? That's the user job these days isn't it?
you get a FREE OS, aka W10, and they turn you into a guinea pig and spy on you and sell your data OR
you get a FREE OS, aka Android, and they spy on you and sell your data OR
you get a PAID "obsolete" OS, aka W7, and they spy on you (a little), sell your data (a little) and you tremble in fear, because FUD, OR
none of the above, aka Apple x-thingy and them linuxes, but hell, neither's for mainstream joe (for different reasons).
Apple does spy, somewhat, but nothing like what other companies do. They don't collect as much information, are better at allowing data to stay on the device rather than getting sent to them, and don't make a large part of their business from data exploitation. So yes, I'm saying that.
How many of us on his forum have a "test device" enrolled in the developer previews? I know I do and apart from periodically updating it and having a prod about I literally do nothing on this device. So as far as Microsoft are concerned the data they slup from it is all good?
How many of us on his forum have a "test device" enrolled in the developer previews? I know I do and apart from periodically updating it and having a prod about I literally do nothing on this device. So as far as Microsoft are concerned the data they slup from it is all good?
I have one like that. Other than using it to buy music off of whyTunes (because Apple can't be arsed to make an Android client or make a way to buy directly off the website), or my daughter borrowing it while I'm fixing her laptop, it doesn't do a lot other than letting me see what crap is coming down the line.
I *HAVE* considered the possibility that MS should make an automatic test/validation suite we could run on our TechPreview machines when we're not doing our own "testing" on them. Sure, it's giving them another excuse to continue not doing their own testing, but that's what their going to do anyway.
Looking at the evidence over the last few updates, apparently many/most do their testing in virtual machines rather than real hardware. Thus, where the HAL hits the hardware, this hasn't been tested against the (literal) myriad of systems out there. A monoculture if ever I've seen one.
Were I even using Windows 10, I can see that I might do the same. Instead, my laptop uses Windows Server 2016. First off, it's only used to connect and do things on the Internet, my canary in the coal mine, not just about Windows but what is "Out There" to get my devices. [All other devices are energy-gapped save the tablets. No big deal restoring.] The other reason is that the laptop can't seem to keep a working Windows 10 alive; has no problem doing the same with Server 2016. Same device driver model, supposedly, for both. Go figure why.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
