[rfc-i] [IAB] draft-iab-xml2rfc-03, "2.19 <displayreference> "

[julian.reschke at gmx.de (Julian Reschke)]

On 2016-03-02 16:42, Joe Hildebrand (jhildebr) wrote:
> On 3/1/16, 11:32 PM, "Julian Reschke" <julian.reschke at gmx.de> wrote:
>
>
>
>>> - Links generated from <relref displayFormat="of" target="EPP" section="2.3"/> (aside from being nonsensical for <referencegroup>) would generate a link to "#STD69", displayed as "[EPP]".  Same with displayFormat="parens"
>>
>> No, they would generate an error. relref/@target needs to reference an
>> @anchor in the document, otherwise the XML is invalid.
>
> My example was inaccurate.  What about:
>
> <relref displayFormat="of" target="STD69" section="2.3"/>
>
> I think that would generate something with [EPP] in it.

I'll change that to an RFC to avoid another topic :-)

<a href="...../rfc69.html#sec-2.3">Section 2</a> of [<a 
href="#RFC69">EPP</a>]

So, yes.

>>> - <relref>'s in OTHER documents probably have to refer to STD69, not EPP.  I'm not sure how they're supposed to figure that out, but hopefully nobody ever does that.
>>
>> Yes. (But why would another document want to reference an entry in a
>> different document's references section?
>
> For example, a document discussing how references should be formatted. :)  If it's not valid to point to any ID in a document using a relref, we should probably give some guidance.

<relref> (which btw should be just a special case of <xref>) was 
designed to point to things that resemble sections. Maybe we have to 
discuss this, but the equivalent feature in rfc2629.xslt has served me 
well so far...

>> It addresses two use cases:
>>
>> a) Citation tags that are not valid as IDs, such as those starting with
>> a digit.
>>
>> b) People including references from other sources, thus loosing control
>> over what citation tag they get.
>>
>> a) is a real use case (although a bit on the edge). For b) I have less
>> sympathy (if including things hurts, don't do it; copy & paste works as
>> well).
>
> It doesn't sound like you're strongly in favor of keeping this functionality.  Are there others who think it's more important than this?

I'm mildly in favor because I already invested time to implement it.

Could you clarify where you see the actual problem?

Best regards, Julian