[gopher] Re: Gopher Protocol Issue

[emanuel at heatdeath organisation <emanuel@xxxxxxxxxxxxx>]

Cameron Kaiser <spectre@xxxxxxxxxxxxxxxxxxxx> wrote:
> I've also been playing with the gopher->web gateway at
> heatdeath.org, which is great stuff. 

I've just joined and I see I've already been mentioned.  That's a
good sign :-).

Cameron Kaiser <spectre@xxxxxxxxxxxxxxxxxxxx> wrote:
> One would think the problem of testing on
> multiple clients would be ameliorated in gopherspace, though. :-/

Well, it's not nearly as bad as with the web.  Sure, you have to test
server software against many clients when developing (and vice-versa),
but putting up a gopher site with established server shouldn't need
testing with all clients.

David Allen <s2mdalle@xxxxxxxxxxxxx> wrote:
> In terms of 'popularizing' gopher again, it seems as if any movement
> towards gopher again (ala Gopher Manifesto) would need to start at the
> browser level, due to the fact that that's what people are addicted to
> at the moment.  But that's just my speculation.

This is why I started my gopher->web gateway (named, suprisingly enough,
webgopher).  It skips the whole problem of having to build good gopher
support into the web browser itself by gatewaying gopher to HTTP and
HTML that will work with any web browser.  This gets especially
important as a trend is to embed web browsers in more and more devices.

Point your gopher to gopher.heatdeath.org and follow the links to see
the gateway in action.  Development is stalled for the moment because
work is getting busy, but I'm in the middle of doing full Gopher+
support.

Cameron Kaiser <spectre@xxxxxxxxxxxxxxxxxxxx> wrote:
> Gopher+ is sort of a nebulous extension and there are a lot of
> inconsistencies in it, the one you mention being one of many. 

Actually I found that after several passes through the Gopher+
"specification" all the inconsistencies I thought I saw in my initial
passes were resolved.

I don't think what was mentioned was an inconsistency at all.  The spec
says that the number of bytes of data must be specified (or -1 or -2).
As you said, UMN gopherd happens to not specify the size, and instead
gets the client to read until the end, but that was simply their
implementation choice.  Yes, having to translate the file as it is sent
makes it more difficult to specify the size, but the server could cache
the actual data size (doesn't UMN gopherd already cache a bunch of
stuff?) so it only needs to calculate it once.

-- 
emanuel at heatdeath organisation
gopher.heatdeath.org