W3C home > Mailing lists > Public > whatwg@whatwg.org > August 2009

[whatwg] Dates BCE

From: Ian Hickson <ian@hixie.ch>
Date: Wed, 5 Aug 2009 02:08:49 +0000 (UTC)
Message-ID: <Pine.LNX.4.62.0908050144090.6420@hixie.dreamhostps.com>

A number of e-mails were sent recently regarding the subject of historical 
dates. This topic was last covered in March in this e-mail:

   http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2009-March/018888.html

In the interests of not rehashing old ground, I have omitted from my reply 
below replies to e-mails that are redundant with the reply that I sent in 
March, cited above.


On Thu, 30 Jul 2009, Sam Kuper wrote:
>
> [...] here are a couple of good examples with ranges:
> 
> http://www.darwinproject.ac.uk/darwinletters/calendar/entry-10762.html
> http://www.darwinproject.ac.uk/darwinletters/calendar/entry-295.html
> http://www.darwinproject.ac.uk/darwinletters/calendar/entry-6611f.html
> 
> Now, either there should be markup available for ranges, or it should at
> least be possible to specify components of a date independently of each
> other, and to imply (at least for humans) a "range" spanning these different
> date elements as appropriate.

It's not clear to me what the use case here is. Without a reason why 
making this data machine-readable is important, it's hard to find good 
ways to solve the problem.

However, it was recently pointed out to me that Microdata is actually a 
quite good way to address this. For example (and I'm making up a 
vocabulary on the fly here, I'm sure a much better job could be done), one 
could cast part of the first of the pages above as follows:

   <span item="com.example.daterange"
    >[<span itemprop="com.example.start" item="com.example.date"
    ><span itemprop="com.example.month">Jan</span
    ><meta itemprop="com.example.year" content="1877"></span
    >-<span itemprop="com.example.end" item="com.example.date"
    ><span itemprop="com.example.month">June</span>
    <span itemprop="com.example.year">1877</span></span>]</span>


On Thu, 30 Jul 2009, David Singer wrote:
> 
> It allows you to build databases with timelines, that span documents on 
> the web from diverse sources.

This is apparently already possible:

   http://newstimeline.googlelabs.com/
   http://www.google.com/search?hl=en&safe=off&rls=en-us&tbo=1&site=mbd&tbs=tl%3A1&q=kittens&aq=f&oq=&aqi=g10


> It would allow you to determine that *this* event reported in an arabic 
> text with a date referring to a caliphate was actually almost certainly 
> *before* this *other* event reported in a byzantine text with a date 
> that is on the indiction cycle.  The experts in arabic and byzantine 
> texts individually might well have the skills to convert these dates to 
> a uniform day-labelling system, whereas the interested reader might have 
> the skills in one or the other, but maybe not both (or perhaps even, 
> neither).

It's unclear to me that there is actually demand to do this on the Web 
from the communities that would be required to actually do it. If there 
was, we would see efforts to work around the limitations in HTML, e.g. 
using Microformats-like solutions. Have we?


On Thu, 30 Jul 2009, Sam Kuper wrote:
> 
> For projects like the Darwin Correspondence Project, machine readable 
> HTML markup of dates might well simplify the various rather fragile and 
> complex custom date search mechanisms these projects have historically 
> tended to use, allowing users to access materials more easily and making 
> APIs to such online corpora easier to create.
>
> Suppose you wanted to mash up the Darwin correspondence data with a 
> SIMILE Timeline[1], it would help if the correspondence data was (more) 
> machine-readable. Now suppose you also wanted to add some diary 
> entries[1] to the same timeline, so that you could instantly visualise 
> when letters were written vs when diary entries were written. This would 
> be much easier if both the two websites from which you were sourcing 
> your data used a consistent, machine-readable date format.
> 
> [1]http://www.simile-widgets.org/timeline/ 
> [2]http://darwin-online.org.uk/content/frameset?itemID=F1925&viewtype=text&pageseq=1

These seem like problems that, if the people involved agree that they 
should in fact try to solve them, could be solved first using microdata or 
Microformats vocabularies.


On Thu, 30 Jul 2009, Mike Shaver wrote:
> 
> I guess it's somehow useful to express this information to machines but 
> not to humans, but sure -- I don't care if it's done with microformats 
> or morse code.  Can someone build and experiment with this to see if 
> it's as useful as we're speculating, and if it requires standard rather 
> than standard-based library support, before we make all browser and 
> other developers bake it into their software?

Indeed.

As is described in the FAQ:

   http://wiki.whatwg.org/wiki/FAQ#Is_there_a_process_for_adding_new_features_to_the_spec.3F

...it is generally preferred if problems can have demonstrated need before 
we add features, especially features as potentially horrifyingly 
complicated as is being discussed here. There are a number of possible 
technologies that could be used to address this; I would recommend taking 
them, e.g. making a microdata vocabulary, and seeing if the relevant 
communities are actually interested in solving these problems that have 
been listed in this thread.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'
Received on Tuesday, 4 August 2009 19:08:49 UTC

This archive was generated by hypermail 2.3.1 : Monday, 13 April 2015 23:08:51 UTC