Slashdot Mirror
Most Expensive JavaScript Ever?
ekran writes "A while ago Opera Software needed more servers. Not just a few servers either — they were planning Opera Mini's growth, implementing Opera Link, and My Opera was also growing quickly.
Most of the major hardware vendors grabbed their specs and came back with offers and sample servers shipped all the way to Oslo for testing. One of the biggest vendors, however, did not do their homework. They shipped the server, but when the Opera sysadmins started up the web-admin interface, they were met with a JavaScript statement that managed to piss off the whole company including the CTO. The script, apparently, locked out the Opera web-browser."
Don't be a jackass. Their market share has nothing to do with the article. I don't care about Opera but the story is still funny as hell and worth repeating. I just wish that kind of stuff happened to me at work. :(
I am the lawn!
Why? I can see not bothering to test much on Opera, or putting a "We only support browsers X and Y." statement in the manual, or even not bothering about any issues that crop up if you use Opera to access the admin interface; but why would you deliberately add a check that breaks Opera?
This nonsense would never happen.
I started as a web developer in the mid-90s. I know how hard it is to develop for multiple browsers and versions. When Netscape and Internet Explorer 4.0 came out, they quickly gained the majority of market share. Many colleagues did not want to keep their sites compatible with 3.x browsers because they felt it was a pain. I would always hear the sentence, "They only have a 5% market share."
To me this was and still is a ridiculous attitude. You're OK randomly raising your middle finger to 1 in 20 potential customers visiting your site? What if that 1 in 20 is the wrong person? Obviously, in this case, they definitely raised their middle finger to the wrong people.
But this gets even worse, because Opera is not obsolete and is fairly standards-compliant. To top it off, the vendor specifically broke the web site for the browser they were too lazy to design for, rather than doing something that makes sense -- like investing time and money to reach a small but tech-savvy segment of the population.
All told -- shamefully -- it makes me feel a little Schadenfreude that it bit them in the rear.
The Institute of Incomplete Research has determined that 9 of out 10
Classic case of a company not knowing what their product is used for.
Opera != Work Browser.
Opera == Bestest P0rn Browser ! Swift image resizing, superior mouse gestures, and remaining responsive even after a gazillion tabs are opened.
It's like turning up with your purpose-built race car at the city center, and whining about speed humps.
Actually, he has a very valid point. In your Toyota example, you are forgetting that in order to support Toyota, you have increased development costs and time in order to support a very, very small percentage of users, compounded by the fact that supported browsers are free and can be installed in a matter of minutes.
If you hire a good developer your site should work for all browsers. It's not fucking rocket science.
Literalism isn't a form of humor, it's you being irritating.
Thanks for that, now I feel kinda like an idiot. At least I am now a smarter idiot than I was half an hour ago.
-Rick
"Most people in the U.S. wouldn't know they live in a tyrannical state if it walked up and grabbed their junk." - MyFirs
You are assuming WAY too much. Some subcontractor wrote software to a specification that said support browsers x and y with functionality a,b,c,..q and pass these tests. To make sure some clever tester didn't find some obscure bug with Opera that would keep him from getting paid he just errored the browser out. This passed functionality testing and was approved. Later this now approved and standard code was delivered with the imbedded management card of a server that just happened to be shipped to Opera. The two events have nothing to do with each other and it unbelievably unlikely that for an account the size of Opera that someone would have done regression testing on their software to make sure it works with Opera's product.
There are 4 boxes to use in the defense of liberty: soap, ballot, jury, ammo. Use in that order. Starting now.
It is common to deliberately add a check that breaks the whole stuff when some 'unexpected' condition happens. You know, assertions.
Which one is better? Not working at all, or seems like working but a not-so-commonly-used-some-sort-of-admin-command somehow gets screwed and the web browser fires a do-not-touch-this-unless-you-want-complete-meltdown-command because there was some minor difference on the javscript engine parsing some parameters? Yeah, can be extremely rare, but if it isn't tested, nobody can be sure.
Obviously, the best thing to do would be to test all possible conditions. However, if you can't, then there can be three choices:
1) Leave it to the users, Nah, I'm not gonna test it.
2) Launch a big warning message and blame the users if something goes wrong, or
3) Make it never work when some unknown condition is reached.
Number 1 is perfectly reasonable when the worst consequence isn't so bad. For example, a web forum interface, or things like Facebook. Maybe number 2 would be better in most cases. But, if an untested scenario may cause huge, irrecoverable damage, number 3 may be the best choice. (You should remember that the product in question was the server management console, which can bring the whole datacenter down when things go wrong.)
My opinion is that, deliberately excluding Opera was a quite reasonable idea. Trying to sell a product that deliberately excludes Opera (web browser) to Opera (the company) was the stupid idea.
You're missing one obvious factor. There are a lot of great browsers written using pre-existing code. Omniweb is a great Mac browser that uses Webkit for HTML, and Spidermonkey for Javascript. The behavior of that javascript is known, but browser detection routines for every minor browser made of major software is ridiculous. You write based on standards, and test in a few major browsers. #2 is the only option. NEVER BLOCK A USER.
Don't forget this was demo hardware from companies responding to a tender. At this point Opera was still evaluating the hardware. Two things seem to have happened:
- the software was implemented with an abort for Opera, either because QA was not done for Opera.
- the vendor didn't appear to know what the primary product of Opera was, and what the browser requirements of their admin interface were.
While the first scenario is bad enough, the second is just unforgivable, since it shows to the customer that the vendor apparently made no attempt to know who their customer was.
Jumpstart the tartan drive.
If you hire a good developer your site should work for all browsers. It's not fucking rocket science.
more to the point, try doing less 'fancy schmancy stuff' and you'll find that simple web forms and UI's work JUST FINE with all browsers. even lynx.
I am not a fulltime web devel (I write software, not web pages) but I have been able to get some web sites up and running that use the form/cgi paradigm and they work across ALL browsers. what's not to work?
oh, you want flash and blinking and 'as you type' stuff that happens?
go elsewhere, then. go write some stupid windows program if you want to 'act that way'.
but that crap does not belong on serious web apps. 'web masters' (what a joke..) have ruined the web with all the blink/fancy crap they try to pull off.
the web was designed so that you would NOT have to check (!) what browser you are sending data to.
why the hell SHOULD a server-side program care how you render data? you've done your job, you tagged paragraphs as paragraphs, lists as lists, images as images. you did what html was designed to do!
just annoys the hell out of me that the web was NOT designed for 'making remote word processors'. it just was never meant for such things. we have twisted this nice interoperable (key term that has lost its meaning, sadly) web into some overly complex beast that 'needs' to know how you render data. or worse, wants to TAKE PART in the font sizing, spacing, colors and so on. what a huge mistake; and its already too late to fix the web at this point; too many idiots are trying to make web programming just another bloated GUI.
in a way, I wish we didn't have such fat pipes or such fast processors. maybe then we'd see a return to sanity and some level of minimalism. that's what web UI's were *supposed* to be. tag the elements and stay the hell away from FORMATING.
--
"It is now safe to switch off your computer."
Except the Cadillac doesn't specifically forbid Ford executives from riding in it, nor does Airborne refuse to deliver to UPS clients.
-Billco, Fnarg.com
You asked for it, ladeees and gentlemen! It may not be that grand masterpiece, Ode to a Small Lump of Green Putty I Found in my Armpit One Midsummer Morning, but I think this little joy from Prostetnic Vogon Jeltz should be enough to warm your hearts:
Oh freddled gruntbuggly, ...[drowned out by moaning and screaming]
Thy micturations are to me
As plurdled gabbleblotchits
On a lurgid bee
That mordiously hath bitled out
Its earted jurtles
Into a rancid festering
Now the jurpling slayjid agrocrustles
Are slurping hagrilly up the axlegrurts
And living glupules frart and slipulate
Like jowling meated liverslime
Groop, I implore thee, my foonting turlingdromes
And hooptiously drangle me
With crinkly bindlewurdles,
Or else I shall rend thee in the gobberwarts with my blurglecruncheon
See if I don't.
Cheers!
"What in the name of Fats Waller is that?"
"A four-foot prune."