My understanding is that it's not about the desktop web client; it's about all the back-office web applications that rely on (shudder) ActiveX controls that expect to run inside IE6 on Win2K or, at best, XP.
Rolling out Firefox or Chrome or IE8 to the desktops would be expensive, but not a budget-buster. Replacing hundreds or thousands of proprietary server-side applications written over the past fifteen years would cost billions and take years, and in the middle of a budget crunch that ain't gonna happen.
(For example: I've heard of one widely-deployed town planning system that requires the users to download a proprietary ActiveX control in order to view planning applications on the desktop ... in PDF format. If you don't run the control, you can't download the PDFs; simple as that. Yes, this kind of proprietary lock-in is insane. Unfortunately, it was procured in the pre-Web 2.0 era when "nobody ever got fired for buying Microsoft" looked a lot more solid than it does today -- and now it's out there in the wild, and replacing it would cost millions.)
>Unfortunately, it was procured in the pre-Web 2.0 era when "nobody ever got fired for buying Microsoft" looked a lot more solid than it does today -- and now it's out there in the wild, and replacing it would cost millions.)
I don't believe it.
I'm sure you can find lots of software consultants with posh cars and big offices to tell you it would cost millions. But really. The PDF is on a file server, all you need to do is link it for the client, how hard can it be.
A decent web developer could probably do it and test it cross browser in a few days - lets say it's a man-month, that's going to be what, $20k absolute top for the project. Even if you had to start from scratch it wouldn't take $millions.
He isn't saying it would cost millions to replace that piece of software. He's saying it would cost millions to replace the hundreds or thousands of ActiveX controls in use, with that being an example of how absurdly pervasive the technology is in the UK government.
>a proprietary ActiveX control in order to view planning applications on the desktop ... in PDF format [...] it's out there in the wild, and replacing it would cost millions
I don't think that's what he is saying. But I'll accept that's what you are saying and ask you for a citation on this pervasive ActiveX control - I can then research it as my sister is in Town Planning here in the UK.
ActiveX controls work in IE8, so they probably wouldn't need to be rewritten or ported. But we'll never know. The UK government is short circuiting the process by asserting just the act of testing would be too burdensome. I call bullshit.
[disclosure: I wrote that blog post, so consider my opinions here highly biased. :-) ]
I've spoken with another developer or three that tell me that, although ActiveX controls remain supported, security policies have changed in some meaningful ways between versions of IE. The result is that your controls that were working fine in IE 6 stop working in later versions of IE. I have not personally confirmed this (and god send that I will never have to).
It wouldn't cost billions if it was allowed to be managed by private industry. This is the kind of BS that corporate developers are frustrated by on a daily basis -- painful, but not difficult, drudgery. The crown's position, as the Sauciers recognized, is completely unjustified. Stasis is the easy way out, and just like no one was ever fired for choosing IBM, almost no one was ever fired for leaving good enough alone.
This coming from the guy in charge of a 15,000 seat enterprise migration from MS Office --> OpenOffice.org (and who was in charge last year of a 18,000 seat migration from Exchange/Outlook to Gmail). I have tons of similar war stories... but yet some how we are still able to get shit done.
If Microsoft (or someone else) built a self-contained "IE6 For Legacy Web Applications" download, that worked like Fluid/Prism (making an individual launcher for each legacy app), allowed a newer IE to live alongside it, and was guaranteed (literally, contractually) to function identically to IE6 proper, most of this problem would go away, I think.
The point of the contractual obligation isn't to assume liability in case they aren't the same, though; it's to force Microsoft (or whoever else) to prove, incontrovertibly, that they are the same, and therefore give the companies no wiggle-room. Without that proof, no company is going to bother either way, no matter how much liability they're spared.
Great.. so they test the apps using the saucelabs tool and discover that it doesn't work in IE8.
They then have to contact whoever wrote the software (do they even know _who wrote_ all the different apps they must have?). The vendor says "Sure, we can make it work with IE8 - pay us this much / day". Or they say "Sorry this app only supports IE6 and look at the 10 year support contract that you signed - IE6 is the only browser we support"
Nice idea from SauceLab but the reality is that testing whether it works in a browser is only the first step and not the biggest hurdle. Fixing it is.
(I've also just noticed that SauceLab is a cloud service - yes they have a solution if your app is behind a firewall - and it says it's as easy as a-b-c. Of course the reality is different. I bet a lot of the apps are firewalled to the extent that they're nowhere near an internet connection and would need network admins to then open it all up. Not really practical given all the effort involved)
I'd venture to guess that not upgrading from IE6 is more costly in terms of overall, sustained economic impact. Sure, it costs the government money to upgrade right now. But multiply the costs of continuing to support IE6 across all their vendors over time and I'd bet that number pales in comparison.
The British government is in the middle of the most savage round of emergency spending cuts since world war two.
They're trying to cut departmental budgets by 25%-40%.
They're trying to reduce the civil service head-count by 30% or thereabouts.
There is no money for replacing anything that isn't a life-or-death emergency. They just axed all new school building and structural maintenance across the entire country. Leaking roofs aren't going to be fixed.
With all due respect, upgrading ten year old computer systems (that still work, after a fashion) is the bottom of their priority tree.
Good points, all. I'm not advocating that they should upgrade. I understand they're in budget slashing mode. Just pointing out that the longview approach would likely be cheaper and that their "more cost effective" statement should probably include a caveat that it's based on short-term cost analysis.
Sauce have a point though, if there are new sites being made then cost of supporting IE6 is going up and up for the new sites, for sure. Is there anything stopping the use of IE6-tab or a gov modified version of it?
I work for a startup that supplies enterprise systems to some very large companies that are still on IE6. In my view, when these companies refuse to upgrade it's a way of pushing (very real and substantial) costs onto us. We're now at the point where our starting position in sales negotiation is that we require IE7, if potential customers don't like it it's a negotiating point. If we really had to we'd support IE6, but we'll look to recover those costs in license/support fees.
Ultimately, this is good for the industry. Until companies like us start attaching a visible cost to IE6 support, these departments will not upgrade.
Solution = chrome frame. Now all their wonderful internal apps will still work. And the rest of the web world can move on.
Solution != selenium in the cloud. Sauce is obviously just kidding around here. Selenium doesn't test your site for you. You need to hire extremely skilled testers to build up a solid selenium test suite.
I complained about this decision myself. Continuing to insist upon using IE6 puts civil servants and the public whose data is entrusted to them at extremely serious risk of being compromised - i.e. data being lost, stolen or altered by unauthorized persons or criminal organizations.
I wish this story would have just been a link to the actual decision. This article is just the news story wrapped in a Sauce Labs ad. I love the idea of Sauce Labs but is a tongue-in-cheek response to an actually interesting and controversial news story worthy of its own HN post?
Well, why really should they upgrade? I mean, ie6 does its job of taking you to the internet which is basically what any browser is supposed to do. So it works finely, so I think they are right to leave it alone. Ie9 is coming up apparently, in no more that 3 years ie8 would be considered rubbish too, so they would have to upgrade again.
Rolling out Firefox or Chrome or IE8 to the desktops would be expensive, but not a budget-buster. Replacing hundreds or thousands of proprietary server-side applications written over the past fifteen years would cost billions and take years, and in the middle of a budget crunch that ain't gonna happen.
(For example: I've heard of one widely-deployed town planning system that requires the users to download a proprietary ActiveX control in order to view planning applications on the desktop ... in PDF format. If you don't run the control, you can't download the PDFs; simple as that. Yes, this kind of proprietary lock-in is insane. Unfortunately, it was procured in the pre-Web 2.0 era when "nobody ever got fired for buying Microsoft" looked a lot more solid than it does today -- and now it's out there in the wild, and replacing it would cost millions.)