[Openstack-docs] Convention on <screen> use

Gauvain Pocentek gauvain.pocentek at objectif-libre.com
Mon Mar 24 14:20:38 UTC 2014


Hey,

Thank you for your answer Anne.

Le 2014-03-24 14:18, Anne Gentle a écrit :
> On Mon, Mar 24, 2014 at 3:26 AM, Gauvain Pocentek
> <gauvain.pocentek at objectif-libre.com> wrote:
> 
>> Le 2014-03-21 14:15, Anne Gentle a écrit :
>> 
>>> On Fri, Mar 21, 2014 at 3:52 AM, Andreas Jaeger <aj at suse.com> wrote:
>>> 
>>>> On 03/20/2014 09:37 PM, Diane Fleming wrote:
>>>> 
>>>>> But I'm willing to go along with the consensus on this. And I 
>>>>> totally
>>>>> agree that we need to have one convention that we all use.
>>>>> 
>>>>> But here's why I like two screens:
>>>>> 
>>>>> 1. You can insert text between the user input and the computer 
>>>>> output,
>>>>> such an explanation of what the user might see in the output.
>>>> 
>>>> You still can do this, I see no problem with that.
>>> 
>>> In the Ops Guide, we standardized on one screen wrapping <userinput>
>>> and <computeroutput> but even so it was causing oddities with their
>>> transforms. 
>>> 
>>> I'd prefer <screen> around each. 
>> 
>> OK. So to sum up, we're (I'm?) not happy with our tools rendering 
>> when using multiple <screen>s, and we have rendering issues when using 
>> a single <screen> too, although this only concerns the ops guide. Did 
>> I understand correctly?
>> 
>> Rendering set aside, I think that one screen is more logical, because 
>> you don't type the command in one terminal, and get the output in 
>> another one. And prompts explicitly define what is to be typed.
>> 
>> To reuse an argument from the basic OS setup discussion that took 
>> place yesterday ([0]), we can probably assume that users know how to 
>> use the command line, and that they will understand what is to be 
>> typed and what is the output.
> 
> No, please don't use that argument again.

OK, this is certainly not the best argument I could have used. Sorry 
about that.
What I was trying to say was that we probably don't have to spell 
everything for the reader, and that using 1 screen makes things clear 
enough (in my humble opinion).

Please let me know if I'm completely out of line here, I'll just shut 
up :)

> 
> I was hoping a DocBook markup expert would give us guidance. I think
> that semantically speaking, <screen> is meant to indicate all that
> someone sees on a screen. To quote from the DocBook reference,
> 
> http://docbook.org/tdg5/en/html/screen.html [1]
> 
> A screen [1] is a verbatim environment for displaying text that the
> user might see on a computer terminal. It is often used to display the
> results of a command.
> 
> Having less specific semantic overtones, screen [1] is often used
> wherever a verbatim presentation is desired, but the semantic
> of programlisting [2] is inappropriate.
> 
>> Let's try to move forward and make a decision about this, this 
>> shouldn't be that hard :)
> 
> It's hard because the semantics aren't specific. I am fine with a
> single <screen> since that is what the user sees. I think we can make
> the output nicer. Feel free to file a bug with the doc-builds tag for
> openstack-manuals.

Reported as https://bugs.launchpad.net/openstack-manuals/+bug/1296731

Gauvain



More information about the Openstack-docs mailing list