Re: [Ref-Links] Re: DOIs used for reference linking

[Mark Doyle <doyle@aps.org>]

Greetings,

Re: the crossposting -- I have asked before whether people would like this  
confined to one list or the other, but no one has chimed in... I'd be happy  
to move to one list or the other. ref-links is most topical, but discuss-doi  
seems to have the most active audience though. Apologies for duplicate  
messages.

On Thu, 08 Apr 1999, "Dr. Guenther Eichhorn" wrote:

[a very nice summary of how ADS works]

I like it! It embodies many of the principles I am espousing.

> The other type of link is more complicated, since we have to
> build the address according to the publishers directory structure.
>
> http://www.idealibrary.com/cgi-bin/links/citation/0019-1035/130/323

While I don't know how idealibrary works, it should be noted that this  
style of URL doesn't necessarily map to a directory structure. The APS link  
manager uses this style, but with single letter identifying keys for the  
fields so that they can appear in any order. Thus the slash is just a  
convenient separator that is part of the usual URL syntax.

Often you seen links with &'s in them to separate the fields and this  
should be avoided at all costs because the & does double-duty as an SGML/XML  
entity identifier. Thus, when embedding links with &'s into HTML, one  
should properly convert the &'s to & and the browser should do the  
translation when requesting the URL. Almost no one does this though and  
those links are susceptible to future problems if an entity like &page;  
becomes common in the online world - then &page= would break (the = sign  
would be interpreted as an entity-terminator because it isn't an  
alphanumeric character). I have seen this happen with some web broswers that  
defined the § to be the section symbol.

Your mirroring solution is also right on target.

Regards,
Mark


------------------------------------------------------
Ref-Links maillist  -  Ref-Links@doi.org
http://www.doi.org/mailman/listinfo/ref-links