IE8 support

[rstacruz]

The page needs an html5shiv. We need to make one without external dependencies, though.

[haileys]

Do any Ruby devs actually use IE8?

[rstacruz]

Yes! Not by choice of course, but because we need to make things that will at least work in IE8.

[haileys]

Fair point.

Couple of questions:

• What missing features in IE8 would need to be shimmed/shivved?
• Is it possible to not use a shiv, but to degrade on IE8 in a way that the page still looks and works ok?

[rstacruz]

Oh it's ridiculously simple! Something like:

var elements = 'section nav header footer audio'.split(" ");
for (var i=0; i<elements.length; ++i) { document.createElement(element); }

(I think this one will actually work as is)

[haileys]

👍

[duncanbeevers]

Could we get away with simply not using html5 tags? I think having semantically-meaningful markup is sort of moot in this case.

[rstacruz]

I'd like that better if @charliesome would agree

[duncanbeevers]

On the other hand, this tool seems primarily useful for rapidly debugging back-end problems. Hopefully people aren't having to do that from ie8. ie8-specific problems seem like they're mostly going to lie in the layout/rendering realm, and possibly in the network transport realm. Problems like this I think are better served by tools like Fiddler/Charles, and browser-based tools seem like they won't be much help.

Given the three choices:

1. No semantic markup
2. html5 shim
3. Scold developers and tell them to stop debugging server-side problems from ie8.

I'd probably just choose number 3, since it requires the fewest code changes.

[haileys]

I am happy with any of those 3 options - what do you think @rstacruz?
On Dec 12, 2012 5:42 AM, "Duncan Beevers" [email protected] wrote:

On the other hand, this tool seems primarily useful for rapidly debugging
back-end problems. Hopefully people aren't having to do that from ie8.
ie8-specific problems seem like they're mostly going to lie in the
layout/rendering realm, and possibly in the network transport realm.
Problems like this I think are better served by tools like Fiddler/Charles,
and browser-based tools seem like they won't be much help.

Given the three choices:

1. No semantic markup
2. html5 shim
3. Scold developers and tell them to stop debugging server-side problems
   from ie8.

I'd probably just choose #3#3,
since it requires the fewest code changes.

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/30#issuecomment-11256397.

[jfelchner]

better_errors is a development tool. The only people this would affect is a dev working at a place of business running Windows XP where installations are locked down. AFAIK that's the only situation where:

• The user cannot install another browser
• The user cannot upgrade IE

Because of that number 3 is my choice. Let's keep development moving forward. On the value-to-time ratio this seems like this issue should be closed and marked wont-fix.