Infrequently Noted

Alex Russell on browsers, standards, and the process of progress.

Comments for Thoughts On A Job Done

Hey Bark,

Honestly, having worked on GCF for years, nothing would shock me about organizational upgrade resistance. I've seen or heard it all, from scared users to cash-strapped libraries to IT systems that have fallen out of contract and can't be touched for fear of breakage. And GCF has been my way of doing something to help them out.

But time moves on.

For the end-users who can't move, luckily we're seeing very broad upgrade campaigns. I show the sorts of things users on legacy browsers experience now in a recent talk:

The end of that talk was a call to arms: it's time for webdevs to start treating legacy browsers like the polluters they are and ask those users, very bluntly, to upgrade. Not force, not require in order to get access to content, but as a friction point to continuing to use outdated software. That's the sort of thing that can be effective at scale with end-users who have some ability to choose; e.g., aren't in locked-down enterprise scenarios.

But back to this billion-dollar org: we've provided the Legacy Browser Support extension:

And for years we've helped these orgs ride their investment in legacy software down the depreciation curve. But there comes a time when our interests part; what's good for those orgs may no longer be good for the overall ecosystem. That GCF is riding into the sunset is yet another signal that they are far, far out of the mainstream and will need to pay increasingly high costs to stay there. And that, taken as a whole, is a good thing. Web developers SHOULD be able to charge (quite a lot) more to build experiences for legacy browsers, and those orgs should be hearing from their vendors that running old, less-secure engines is asking for trouble. For organizations with IT staff, this really is the end of the road and they really do have to choose. I know it's painful to hear, and I've done my level best to ensure that these sorts of organizations can't punish the rest of the world by driving clunkers, but now the water has receeded and we can see legacy browsers for what they are. Time to act. And the good news is that these are people with money, so they can buy options, even if they can't buy any more time.

by alex at

Hi Bark,

Thanks for the thoughtful response.

I've worked in organizations running yesterday's technology tomorrow. My first "real" web job was at a place that staffed a full floor at their global HQ with COBOL and JCL programmers. But that was a business choice. As is the situation you're describing.

That the costs of doing nothing inside these organizations is rising is a long-delayed mirror of the costs they've been able to externalize until now. For the first copule of years of IE 6/7/8 ownership, these organizations were in the mainstream, not costing anyone anything, and their costs of ownership were low. As time passed, their collective market pull distorted the cost curve: the entire platform was held back by the drag they created, apps suffered, and they continued to not pay any marginal price for that distortion.

What's happening now is that these chickens are coming home to roost for real. After nearly a decade of making life harder and more expensive for everyone else, the rest of the web has finally left them well and truly behind. GCF was a land-bridge to help them navigate their way to a better world. That they haven't is still a deliberate choice, and one they'll now have to do something about. That's not optimal for them, but we're at the marginal limit of my sympathy: I only care about the choices those organizations made to the extent that they held the larger web back. They have agency, no matter how much they want to talk themselves out of it via budgeting exercises. I have poured years of my life into helping them out, and Google has thrown huge amounts of money into the effort; all of which is justified only on the hope of a better web. Let me be blunt: it doesn't matter the cause of the web evolving what these organizations choose to do now. They can join the modern, evolving web and once again have low ongoing ownership costs or they can ride a legacy platform and pay full-freight. Either way, developers and consumers of apps built today are no longer caught in the crossfire.

Another point: the upgrade cycle you're implicitly positing here is broken. The old-skool model of "lets adopt something and not update it and patch it only occasionally" doesn't match either today's malware threat landscape or the way the rest of the web is moving. Organizations with that mindset helped lock developers into the idea of a stable software target (vs. a stable core set of compatible APIs) and that has ALSO cost us large in the collective sense. We're finally breaking free of that, and even MSFT is hopefully going to move that way with WIndows Blue. The web is one of the platforms that's evolving in a compatible way that ensures that apps built today work tomorrow, regardless of vendor.

If you lose that battle about Silverlight/Flash, you're just going to be betting on legacy. They're both dead-platforms-walking. The alternative to that is to bite off a truly modern model; one where software compatibility is the question, not version number. Updates MUST happen, else you're just asking to have malware-riddled boxes. In today's environment, it's reasonable to talk about alternative vendors for a comodity platform, but weighing versions against each other is nonsensical. There's only "up-to-date" and "pwn'd". Even MSFT and Adobe will tell you that about Siverlight and Flash, respectively.

Lastly, GCF will still be deployable. It just won't be updated, although I anticipate that folks will be making updated open-source installers available. There's some movement on the mailing list about that. But that will get increasingly hard as the patchset gets larger. We've provided a 6 month extrication timeframe to help ease the burden, and I suppose we could talk on the list about alternative timeframes that you think are more reasonable. Any evidence you can provide would be helpful in those discussions.

Lastly, if someone wants to pay one of the GCF hackers huge piles of cash to quit Google and maintain open-source GCF patches against Chromium...well...can't hurt to offer, right?


by alex at
Why not set the end date to April 2014, the end of support date for XP?
Clearly you've never done work for federal, state, or local government.
by Yep at

My wife is front-end lead for I'm intimately familiar with the constraints the slowest users place on the ecosystem. If I wasn't, I'd never have poured years of my life into GCF. But the left-behind eventually need to own their situation. I've done my bit, and while I feel for those trapped in situations beyond their means to effect change, I hope this will be one more signal that the time to move into the mid '00s is now.

by alex at
Alex, great work on GCF, thanks to your hard work, GCF has given us developers much needed relief when client corporate standard browser put us years behind the mainstream, now I know there are sacrifices involved and can never thank you guys enough
by venkat at
Agree with comment above... You'd be amazed and dismayed at the number of Fortune 500 companies still running IE8, with indefinite plans for upgrading. The billion dollar company I work for and develop software for made GCF a part of our strategy for addressing these companies and providing them with the best web apps we could. Now, with Google retiring GCF, we are in quite the quandary. It's a little shocking, given Google's endeavors in the business space, how unaware they seem to be of the state of IT in the country's largest companies.
by Bark at
Alex -- thanks for taking the time for a clearly well-thought response. I agree with your intent, but not necessarily with the justification or approach.

With the latter, the idea that these companies are "people with money" is far too simplistic a view. First, you have IT organizations that are cash-strapped. These are seen as cost centers in corporate America, not revenue-generating think tanks like you might see at the halls of Google and Silicon Valley. Many of them, in the largest companies, are seeing their budgets slashed, their personnel outsourced, and their ability to move forward from a technology perspective is eliminated. They're going to spend money buying their way out of performance and storage concerns, not improving architectural elements of their application deployment strategy.

At one of our largest customers, a company with 61 billion -- billion -- in revenue and a market cap of over 230 billion, their head of IT who I met with to discuss mobile strategy said, without a smile, "at our company, we have yesterday's technology tomorrow." Again, no smile.

The money is not there for IT. This is endemic.

The actual business customers at these clients that I service are just as budget strapped -- the insights we provide come at a hefty price, and it is well worth it when our assets are leveraged. However, to now have to say, after selling in the idea of deploying GCF, that they have to abandon these plans and shift to another browser installation and another set of policies, as well as (believe it or not!) another set of trainings on the new browser... Lets just say that I'm not relishing these discussions, and I generally like a good fight.

Finally, as a technologist, I can understand and appreciate the holistic intent -- ditch the crutch, and users will be forced to make the move. However, if you are really going to be true to your ethic, that you want the best of the web available to everyone, I would advise that you are doing it too soon; that the corporate world is not yet ready, with their legions of IE7 and IE8 installs; and that GCF is still the vector that development companies, such as mine, who have and continue to invest heavily in the vision of deploying the best web solutions possible using the most contemporary technologies, hung our hats on.

And now, regardless of your grand intent (I say that with zero sarcasm), these customers will not get the best of the web. And while it may be premature of me to say so, and while I'm confident that I could win the argument, the issue will be raised at our next strategic discovery sessions, and I will find myself defending my position to boost HTML5 in lieu of our Silverlight and Flash strategy which was entrenched for so long.

Imagine if I lose that battle!

by Bark at