[gopher] What's my IP address in gopher!

[Kim Holviala]

On May 5, 2012, at 15:07 , Denis Bernard wrote:

> Kim Holviala <kim at ...> writes:
>> 
>> I rule!
>> 
> 
>  Your idea to provide IP address with host name and country is good... but
> implementation  is not compliant with rfc-1436 as you are using the “i” item
> type.
> 
>  You should generate your  message in a specific text file, hence using a “0”
> item type.

No need for that if I ever wanted to stuff with type "0". The server fully supports CGI's so it's just a matter of writing a script... Even autodetection between types 0 and 1 works farly well so the same script could just choose which format to output.

>  I do not like “i” item-type, not only because it is not strictly compliant to
> the standard, but because it is a bad idea for the small devices (like smart
> phones).

The strict standard alse defines that all transfers end with a ".", but only very vaguely (it's been a while since I last read the RFC). If my memory serves me correctly the client just has to guess whether to remove the last dot or not...

> Having nice ascii-arts is not good for the readability of menus for this small 
> displays! Making drawings in the menus is going in the same way that the Web
> sites!

This I have to disagree, strongly. ASCII art has been with us 20 years before anyone even dreamed about graphics. It's old, just as old as terminal displays and hence I see no problem using it.

Now, if the client decides to do "smart" reflow messing up the art then it's a client problem, not a server problem.

>  Run this to have a look at your gophersite:
> 
> www gopher://gophernicus.org/1/
> 
> then, my own gophersite: 
> 
> www gopher://oceamer.com/1/

As a counter-argument my site has no charset problems while yours looks like "événement". If you're optimizing for small devices then you cannot use anything else than 7bit US-ASCII. In fact, if you Obey The Law(tm) (well, the RFC) then you cannot use anything else either because it doesn't define any charsets to use. Historically everyone used Latin-1 so that's a good approximation, but that breaks with modern UTF-8 terminals.



- Kim