50 comments

[ 3.3 ms ] story [ 71.5 ms ] thread
Thank god. There are still people in my office using XP and IE 6/7 that I wonder why they cant do or see certain things.
I can understand being on Windows XP (even though it’s still ridiculous) but what possible reason is there not to upgrade to IE8?
There should be a FAQ; business critical process is only supported when accessed with IE6, regulatory/licensing limits you to a certified configuration which specifically demands IE6, etc.
Either the user refuses to upgrade or the higher ups determine its more cost effective to use IE6 instead of upgrading the Application itself.
XP supports up to IE8; why on earth haven't they (been) upgraded?
I would imagine someone went a bit wild on the ActiveX components in IE and needs them for their custom apps to work.
We have lots of people in the field doing service related work all over the world and they do not take the time to upgrade (or do their job but thats another story). Its takes 2-3 months just to get the old computer back after we send then a brand new computer.
Web apps that only officially support older versions of IE. In some cases, blocking entirely other browsers even if they would work. And even when they don't, the vendor only officially supports certain browsers and rejects any incidents logged using other ones.

The exact same thing is a key reason why XP has remained in use for so long.

This is what VMs are for. If you really really absolutely must run one piece of ancient software, it doesn't need to hold back everything else you do.
XP in your VM is still EOL. Doesn't really help you. Also, you double your licensing costs.
> XP in your VM is still EOL

Doesn't matter that much if it's on a private network.

> Also, you double your licensing costs.

Not if the OS and virtualization solution require no license.

For you or I that is an easy solution.

For a large organisation IT/TS cost is a problem with that solution. As well as managing everything else about their user's machines they'd have an extra app/service to maintain (whichever VM solution they go for) and an extra OS (XP in the VM) to somehow keep up-to-date for each user. Not insurmountable by any means, but far from zero hassle/cost.

Once an ancient version of IE is out of support, assuming that is the app you are running in a VM in this case, you also have an extra complication: making sure the VMs can only talk to your intranet sites that require it, as you don't want your users browsing more public hosts with a potentially insecure UA. Again not rocket science, just some routing/firewall tweaks, but something that has to be added to the support cost of your environments.

Depending on the licensing deal the company has with MS (or relevant distributor) there might also be extra cost involved in providing a second OS instance to all your users.

Try explaining what a VMs and why they are needing to use it to C-Level or VPs that dont understand technology.
I have even seen old webapps that are handled better with newer browsers be restriced to the old versions of IE because of "security reasons."
With XP finally at the end of life, this is a great opportunity for many companies or institutions to switch to Linux.
I'm all for Linux, but this won't happen. The people who leave XP due to running out of paid support options will find no comfort in Linux. Commercial support is too fragmented and not as discoverable as on Windows. People who do not care will stay on XP still. Companies will upgrade to Windows 8, or perhaps 7. I don't see a big demographic switching to Linux right now.
"The people who leave XP due to running out of paid support options will find no comfort in Linux."

I can't be bothered to post links to every distro and every consultant on the planet, but just for laughs:

http://www.debian.org/consultants/

And this is just independents (mostly) and just Debian. I'm sure IBM and other major corps would be glad to help as long as its legal, which is not an issue with FOSS unlike windows.

Now if MS opensourced windows XP as part of the abandonment, that might be interesting.

I just checked your link for consultants in my country. One is a dude with a hotmail address and no website. One is a dude with a simple static website with his picture,name and a message that says basically "We do Everything! Please call me!". One is apparently going out of business. One gave me a warning about dodgy certificates. And finally one that sounded like a legitimate company. Not really something I'd want to bet my company on.
"Not really something I'd want to bet my company on."

The irony of that line in a discussion primarily about microsoft products is fairly humorous.

You're lucky you have anyone in your country :) South Africa isn't listed
> Commercial support is too fragmented and not as discoverable as on Windows.

Ubuntu, Suse, and RHEL are commercial support. Novel, Canonical, and Red Hat make good chunks of profits from support contracts. They have a business model at all because of businesses that adopt them and pay for support, and Red Hat is a billion dollar business!

Note most of that support is server side, but I know all 3 offer some variant of standard desktop support as well. So do a lot of Ubuntu PC vendors like system76.

You're also forgetting re-training thousands of employees on an entirely new operating system. That's far more of an issue than commercial support at this point (as the argumentors responding to you have pointed out).

Even though Unity and other desktop environments are quite intuitive at this point, changing over would cause a massive disruption to laymen who have been using a Start menu, Windows Explorer, and Outlook for years.

The benefits of switching (license costs, I suppose) far, far outweigh the benefits.

Or macs, or one of a zillion tablets.

The most entertaining part to watch is psuedo-embedded devices. All you need to do to upgrade, is throw out an entire $10M MRI machine and buy another. Yes your steam boiler is OK but the monitoring system runs XP and we don't sell upgrade kits of just a new monitoring system, although we could make you a deal on a completely new steam boiler system. Or a (in)security system, a HVAC system that runs independently but can only be reprogrammed using some weird XP only driver and software, or a zillion other PC assisted technologies.

The kind of company that is still on XP and is switching because it is EOL isn't the kind of company who will be switching to Linux any time soon.
> According to Microsoft, a full operating system migration can take up to six months

Well, we have to migrate about 6000 pcs, and it's a 1 and ½ years project. Much more than six moths.

So why did you start so late / why does it take so long?
Because XP probably does everything they need and the upgrade will be reluctant.
A lot of businesses I talk to have a lot of legacy applications which work well under XP, but break for whatever esoteric reason (for example, an old Access database, or VB6 app) under Windows 7/8. An equally large pain the ass is a ton of intranet sites break when switching to IE9 and above (heck, a lot of companies still force IE6/7 and below with Group Policy). It's difficult to make a case to management when everything is working well (it's hard to explain the nuances of "end of support life cycle" to someone who needs help "getting to the Google").

Also, there are a fair number of desktop software products which don't even support Windows 8 yet. Hard as it may seem for the crowd here, most of the software world is not on the HTML5/Ruby/OMGMVCJAVASCRIPTBRO bandwagon yet.

I can't speak to the websites having problems with IE9 and above, however the desktop application compatibility situation has ended up being overblown in my experience. Sitting down with "Process Monitor" and weakening permissions or spending a few minutes figuring out how to apply the existing Application Compatibility shims the OS provides has sorted out almost every problematic desktop application. For the few applications that failed to yield the "Windows XP Mode" VM (which comes with Windows 7 Professional licenses) has worked fine (albeit the multiple users per PC "story" for "Windows XP Mode" is terrible and having a bunch of XP VMs to support adds complexity).
What's wrong with the OMJB framework? I've written hundreds of large-scale multi-optional dynamic polymorphic backend front-end GUI shell's on legacy mainframe infrastructure with it.
(comment deleted)
Making sure all the software and custom apps work with conversion of those that don't takes a fair amount of time in a big organization. I remember one place going from NT 3.51 to NT 4.0 and all the fun of the "new interface" back then.
I have a Windows XP VM on my computer that runs my client connection to a piece of essential hospital software. It cannot be run on a machine that has .NET 3.5.1 and > 4.0 installed simultaneously, or it throws assembly compatibility errors. The ActiveX components seemingly only work in IE8, not even IE9 in compatibility mode. The vendor will probably only fix this compatibility problem when the piece of software has even less support than VB6 (i.e., virtually none). Because healthcare.
"Windows 7 or 8?"

By the time they've made that decision, Windows 9 may be out :)

The big thing about the end of XP is the non-support of SNI (server name indication) in TLS (SSL) in any version of IE on XP.
Wow. The best reason to upgrade is to remove the version lock on the default browser. Not sure what all that says about Microsoft and the current state of technology, but it is weird and sad.
I don't get this, though. The default browser before IE10 was awful, so you shouldn't use it anyway.

It isn't my fault you have incompetent higher ups that seem to trust MS with their lives but won't even consider Google or Mozilla.

You've obviously never developed for a major corporate customer. If I tell some of our large clients "I'm not supporting IE8 and below because it is crap and you shouldn't be using it (or you could install something better alongside if you really need old IE for legacy code)" they will respond with "fine, we find someone who will support our current standard environment then, have a nice day".

We as devs and server/service admins don't get the luxury of dictating what browsers our customers use, and insulting our own higher-ups will do nothing to change that as it is out of their control too.

As much as everyone at my employer would love to drop support for IE8 and lower, it would save a pile of QA time and front-end developer frustration, we don't want to lose lucrative contracts with large investment banks in order to have that luxury and that would happen at this point in time if we dug our heals in on the matter.

Welcome to the real world. I'm sorry to say, but it isn't always nice.

I do corperate IT work. Firefox & Chrome cannot be managed while IE is integrated with Active Directory. Honestly it is a pain when the end user gets tricked into installing Chrome and now none of their work websites function due to needing a real Microsoft Active X runtime.

Your computer won't boot/Windows corrupted? Sorry you just lost all your history & bookmarks if you used IE I could have transfered them to your new computer for you (IE stores them in a folder, Firefox you can but it is a pain, Chrome you are hosed.)

*corporate

IE is not integrated with Active Directory.

IE settings can be controlled through group policies, and there are policies that can be applied to IE that MS bundles with Windows Server.

Firefox and Chrome can be controlled in a similar manner through group policy by downloading the administrative templates from the respective organizations.

On chrome you are not "hosed". All of the bookmarks are stored in a single file aptly named "bookmarks", just move the file to the new computer. Done.

Firefox also stores your bookmarks in a single file. Just replace the file on the new computer and you're done.

Surely copying a single file can't be more of a pain than copying many individual files.

Nobody uses ActiveX on the web anymore, due to it's IE-only nature. Sure, websites still exist, but they have become rarer over the past half-decade.

With your example: if you have a user that can't figure out how to click on a different browser icon on their desktop, this is nothing more than a user training issue.. or if you are that adamant that Chrome is a serious problem for you, just use group policy to prevent your users from installing it.

I manage several thousand users, most use chrome, some use firefox, and internet explorer is also there for people that aren't familiar with other browsers, or need to access special (poorly designed) websites. All browsers are managed with group policy and bookmarks are managed with ease.

User complaints are at zero because so far everyone has been able to tell the difference between the icons.

It isn't so much a reason to upgrade as a benefit for devs and server admins to be thankful users are upgrading. It means we can start considering SNI for significant deployments. SNI has worked since IE7 if used on Vista or later.

The best reason to upgrade is that as of April next year XP is no longer going to get security updates unless when a problem is found you pay MS to release a fix for it.

Our major clients (two of the largest investment banks in the country) have recently moved to IE8 on XP and will soon be moving to IE8 on Win7. They will be keeping IE8 as they have a lot of legacy code that they don't want to spend time/money testing (and fixing as needed) against browsers (including IE9+) that don't share all "classic IE"'s bugs upon which some code may be accidentally dependent. See http://stackoverflow.com/questions/17671383/ for an example where newer IE has trouble with code that worked in older versions - there are examples that don't involve ActiveX too so even if ActiveX isn't an issue (it is for some of our client's legacy "intranet" apps) you could hit problems like this.

The only reason they finally moved off IE6 is that is soon drops out of any support window even for security updates (this is the main reason they will soon be moving off XP too, the other biggie here being new hardware increasingly not having drivers available that work with XP). Win7 and IE8 leave their final extended support window in 2020 IIRC so we'll be stuck supporting IE8 for some time, but we can at least consider using SNI once we know they've dropped XP on their standard machine builds.

Of course anything that targets mobile people might still not be able to consider SNI even once XP is long enough buried, as some mobile browsers won't play ball. There are still a lot of Android 2.x devices out there and the stock browsers on those won't like it for a start.

Seriously. I was scratching my head for 2 days (!) trying to figure our why some users on IE8 couldn't access our CDN. Turns out it was this SNI issue that I had no clue existed.

We had to create a separate machine as a CDN load balancer specifically to handle XP.

If you're migrating off XP you should equally consider migrating to Linux.
Yes, because it's not like any of the applications you use on XP will work on Windows 7 or 8.

And Exchange support is so wonderful on Linux apps. And Active Directory, that's just a walk in the park.

Microsoft made it way too hard to upgrade out of XP (unless you updated through Vista.) It is no wonder everyone is still on XP.

It would light the PC industry on fire if they made it one click to move your stuff to another machine.

You can buy a new Mac every year and only wait a few hours to move everything, but on Windows it takes months.

That could be a bit of an apples to oranges comparison. From what I've seen the migration time for large corporate XP deployments has little to do with the OS upgrade itself and everything to do with the application support issues.

I was at a meeting recently where people were still saying that they had applications which would only run on XP, which was why they hadn't upgraded.

The problem for corporates now is likely that they won't/can't complete an upgrade before the deadline and the risks of having unsupported XP are unknown. One theory is that a lot of potential attacker will be holding any XP exploits they have at the moment with the intent to use them after the end of support, when there will likely be no patch.

No. Going to new Apple releases has the same problem: app compatibility, and OS X doesn't support stuff going back to the 90's, either. And that's just for small personal apps on my iMac, let alone bespoke applications on windows that the Enterprise rely on.
1. Switch to an app with a backend that can talk to a tablet and a browser.
Not everyone is using software in a standard office environment. I have a CNC machine running Windows NT 4. Its a production machine; I need it running, period. I have another running Windows 2000; its a tooling machine, also essential. I have PCs running various CAD and CAM programs which cost 5 figures, and are not negotiable; I simply need them at all times. I need my server of all our past work available at all times; I need the network to be equally available. I can't afford the disruption of trying to get all this running on a new OS because Microsoft thinks I have nothing better to do. I have product to get out the door.