[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: [gopher] gopher++ (gopher1) protocol



On 2010-01-11 15:49, Kacper Gutowski wrote:

Yes, servers would be *a lot more* complicated. But clients would be
simpler as they could count on the server to do what's being asked.

In order to support your Gopher++, clients would need to have
sophisticated detection and fallback routines because server is
free to fail some conversions even if it supports Gopher++.

Well, everything falls back to gopher0. So basically, what you're saying is that "rfc1436"-compliant gopher client needs sophisticated detection and fallback routines - which incidently is true.

[image transcoding]
I would choose to do a plain text. After all that's what is Gopher for.

True. I got a bit carried away with stuff when I was writing it... I think I should remove all traces of audio and video from the doc. Even the image format conversions are a bit off - but definately useful as images *are* important.

And if there are really such strict limitations I'd simply stick to
original Gopher protocol.  Mind that many would like to put their servers
on such limited devices.

100% backward and forward compatibility. What's stopping you from using a text-only gopher0 client? Why would you be bothered at the extra stuff the server offers?

do with text that has high-bit characters? Do you count on it being
Latin-1 like the original RFC says? Do you trust it to be UTF-8? Do
you try to magically sniff the content charset?

I would be really happy to be able to just assume that's UTF-8.

You can't, as gopher0 doesn't say it's UTF-8.

Server can not do much with non-Latin text when asked to serve it
as US-ASCII anyway.

If I serve out standard Latin-1 menus my Firefox (both Win and Linux) barfs and won't print those resource lines (don't know why). Stripping the high-bit chars away at least lets me read most of the text. This isn't some hypothetical thing either, this is what I encountered with my UTF-8 filenames I was serving out.

Seems pretty good to me.... If we take 1 server and 100 clients, I
know where I would prefer my file format support to be - in the same
place where the actual files are.

That's a 1 server going out of resources pretty fast.

That might be, but we don't know it until someone (me) tries it. But I don't think the server runs out of resources - I'm not planning on doing the conversions more than one (caching does exist).



- Kim




_______________________________________________
Gopher-Project mailing list
Gopher-Project@lists.alioth.debian.org
http://lists.alioth.debian.org/mailman/listinfo/gopher-project




Reply to: