[tdwg-guid] Embedding LSID links within Publications

Fri Nov 30 22:07:30 CET 2007

Ricardo's email prompted me to launch into the next round of questions
concerning the implementation of LSIDs in real-world taxonomy.  His comment
was this:

> In short, the document 
> recommends that you use the LSID in its pure form as 
> the link text and the proxy version as the link URL.

I have long believed that GUIDs (and especially LSIDs) were meant for use by
computers, not humans.  As such, I am generally opposed to the idea of
printing or displaying LSIDs in a way intended for humans to read them.
They should be hidden, behind the scenes, enebling the cross-links.

In the case of the five new species of Chromis, I think it's appropriate to
display the ID's in the human-readable parts of the document, because these
are technically the ZooBank Registration IDs for the nomencatural acts.

However, here is where the first question comes in.  As I've described
previously, my strategy with ZooBank is to assign UUIDs to the records, and
then use the UUIDs as the "ObjectID" part of the LSID, such as:

urn:lsid:zoobank.org:act:20889795-7EC7-42F3-A4C3-D1D97704A609

The idea behind using UUIDs is to "hedge my bets" agaisnt the possible
demise of the LSID standard.  In this sense, the LSID syntaxt is really just
a wrapper representing a resolution protocol for the "real" GUID, which in
this case is the UUID.

So....if I am to print/display the "ZooBank Registration ID", what should it
be?  The full LSID (which is placing a bet that LSIDs will still be
recognizable/resolvable 250 years from now); or just the UUID (which is the
"true" GUID for the ZooBank record, independent of the mechanism used to
resolve it)?

Let's assume, based on the other thread discussion, that I will be wrapping
the LSID within an HTTP proxy to make it clickable for the URL part of the
hyperlink.  The question is, what do I show the human reader?  Options are:

<a
href="http://zoobank.org/urn:lsid:zoobank.org:act:20889795-7EC7-42F3-A4C3-D1
D97704A609">20889795-7EC7-42F3-A4C3-D1D97704A609</a>

OR: 

<a
href="http://zoobank.org/urn:lsid:zoobank.org:act:20889795-7EC7-42F3-A4C3-D1
D97704A609">urn:lsid:zoobank.org:act:20889795-7EC7-42F3-A4C3-D1D97704A609</a
><a href="http://lsids.sourceforge.net/"><img
src="http://zoobank.bishopmuseum.org/images/lsidlogo.jpg"></a>

...the main difference being whether the human sees the "ZooBank
Registration ID" (UUID), or sees the "ZooBank LSID" (=ZooBank RegistrationID
wrapped in LSID syntax)?

I'm torn on this -- part of me wants to keep it to just the UUID (which
makes the publication look slightly less stupid 250 years from now, in the
event that LSIDs do not last more than a few years); but the other part of
me wants to trumpet the notion of LSIDs to support the general TDWG cause,
and hence set an example of representing ZooBank Ids as LSIDs, rather than
stripped-down UUIDs.

Thoughts?

This also sets me up for the next related question:

Though I can see the justification for displaying, in human readable form,
an LSID/UUID that serves as a "ZooBank Registration ID" within the context
of the actual pblication that "creates" the identified nomenclatural
act....I'm not expecially inclined to sprinkle such cumbersome LSIDs in
human-readible form throughout the entire document.

For example, consider this sentence from the Introduction of the Chromis MS:

"The pomacentrid genus Chromis Cuvier 1814 (type species Sparus chromis
Linnaeus 1758) is the largest genus of the family, with 86 valid species."

One way to mark this up would be as follows:

"The pomacentrid genus <i><a
href="http://zoobank.org/urn:lsid:zoobank.org:act:B8F9F80D-5798-4342-810D-D8
274164F8F1">Chromis</a></i> Cuvier 1814 (type species <i><a
href="http://zoobank.org/urn:lsid:zoobank.org:act:1A66BAE9-9B37-4C73-A560-BF
63D0345F04">Sparus</a></i> <i><a
href="http://zoobank.org/urn:lsid:zoobank.org:act:8F966156-68DE-4147-953D-74
E1B3724F5F">chromis</a></i> Linnaeus 1758) is the largest genus of the
family, with 86 valid species."

This would appear to the user exactly the same as the above quote, except
the three taxon names would be clickable links to the repsective ZooBank
records.

The alternative would be to do something like this:

"The pomacentrid genus <i>Chromis</i> [<a
href="http://zoobank.org/urn:lsid:zoobank.org:act:B8F9F80D-5798-4342-810D-D8
274164F8F1">ZooBank: B8F9F80D-5798-4342-810D-D8274164F8F1</a>] Cuvier 1814
(type species <i>Sparus</i> [<a
href="http://zoobank.org/urn:lsid:zoobank.org:act:1A66BAE9-9B37-4C73-A560-BF
63D0345F04">ZooBank: 1A66BAE9-9B37-4C73-A560-BF63D0345F04</a>]
<i>chromis</i> [<a
href="http://zoobank.org/urn:lsid:zoobank.org:act:8F966156-68DE-4147-953D-74
E1B3724F5F">ZooBank: 8F966156-68DE-4147-953D-74E1B3724F5F</a>] Linnaeus
1758) is the largest genus of the family, with 86 valid species."

To the human reader, this would appear as follows:

"The pomacentrid genus Chromis [ZooBank:
B8F9F80D-5798-4342-810D-D8274164F8F1] Cuvier 1814 (type species Sparus
[ZooBank: 1A66BAE9-9B37-4C73-A560-BF63D0345F04] chromis [ZooBank:
8F966156-68DE-4147-953D-74E1B3724F5F] Linnaeus 1758) is the largest genus of
the family, with 86 valid species."

Kinda klunky, if you ask me -- even moreso if the full LSID syntax is
included -- and more klunky still if I add the little "LSID" icon logos to
each one.

My preference would be to go with the first option, and simply make the text
of the names the clickable part, and have the UUID/LSID hidden from the user
(but still visible to the computer) via the embedded HTTP proxy link.

But if I go that route this leads to another problem:

Because ICZN still requires paper-based publication for the new names to be
available, there will also be a paper version published concurrently
(standatd Zootaxa practice).  If the GUIDs are not visible to the human in
the PDF, they will not be visible at all in the paper copy.

So....I'm now considering adding an extra section to the published version
of the article (both paper and PDF), that comes after the "Literature Cited"
section, and represents a series of "end notes", indexed from the text via
superscript numbers, that display in human readible form the complete URL of
the link.

For example, something along the lines of the following:

"The pomacentrid genus <i>Chromis </i> <sup><a
href="http://zoobank.org/urn:lsid:zoobank.org:act:B8F9F80D-5798-4342-810D-D8
274164F8F1">1</a></sup> Cuvier 1814 (type species <i>Sparus</i> <sup><a
href="http://zoobank.org/urn:lsid:zoobank.org:act:1A66BAE9-9B37-4C73-A560-BF
63D0345F04">2</a></sup> <i>chromis</i> <sup><a
href="http://zoobank.org/urn:lsid:zoobank.org:act:8F966156-68DE-4147-953D-74
E1B3724F5F">chromis</a></sup> Linnaeus 1758) is the largest genus of the
family, with 86 valid species."

Which would appear to the user simply as:

"The pomacentrid genus Chromis^1 Cuvier 1814 (type species Sparus^2
chromis^3 Linnaeus 1758) is the largest genus of the family, with 86 valid
species."

(where "^" indicates superscripted character)

The "1", "2" and "3" would be the clickale links to the online resources,
but at the end of the article there would ba a section called "Linked
Resources" or something, along the lines of the following:

Linked Resources:

1.
http://zoobank.org/urn:lsid:zoobank.org:act:B8F9F80D-5798-4342-810D-D8274164
F8F1
2.
http://zoobank.org/urn:lsid:zoobank.org:act:1A66BAE9-9B37-4C73-A560-BF63D034
5F04
3.
http://zoobank.org/urn:lsid:zoobank.org:act:8F966156-68DE-4147-953D-74E1B372
4F5F

That way, the reader of the paper copy of the publication could, if they
wanted to, manually type the various links into a browser.

This seems like the best compromise solution to me; but there is one
problem:

If the superscript "1", "2", "3" etc are the clicable links, that makes for
a very small tarket.  People not so dexterous with the mouse control might
be annoyed trying to click on the tiny little number.  So I'm thinking maybe
a little Icon similar to the LSID icon
(http://zoobank.bishopmuseum.org/images/lsidlogo.jpg), except numbered.
That would make for an easier mouse target to click on, and also would
distinguish it from other superscript numbers (footnotes, etc.).
Alternatively, I could simply make the superscript hyperlinked text be:
"Link: 1", "Link: 2" and so on.

Well... That's probably enough for now.  I still want to respond to Chuck's
email, but I have to get to a meeting now.

Aloha,
Rich

P.S. Later today I'll create a web version of this, so make it a bit more
intelligible.

Richard L. Pyle, PhD
Database Coordinator for Natural Sciences
  and Associate Zoologist in Ichthyology
Department of Natural Sciences, Bishop Museum
1525 Bernice St., Honolulu, HI 96817
Ph: (808)848-4115, Fax: (808)847-8252
email: deepreef at bishopmuseum.org
http://hbs.bishopmuseum.org/staff/pylerichard.html