[uf-discuss] hRecipe - status quo - how many elements

[Thomas Loertsch]

Hi again,

I'm indulging in hRecipe optimization and advancement today...

Early this year, when hRecipe had just reached Draft-status, Tantek raised
his head and cautioned that the format may contain too many elements and
that it's a *principle* of microformats to start (and maybe stay) small. I
tend more to the opinion that "small" is not a good criterium per se but
that the element set size should be "just right" to be most useful. We
settled the dispute (in which no one else joined) by marking the majority of
the properties as 'experimental' and cautioning their use.

Just as a reminder, this is the element set:
    fn    
    ingredient (value, type)
    yield 
    instructions  
    duration 
    photo
    summary
    author
    published
    nutrition (value, type)
    tag

I checked the usage of the hRecipe format today and was delighted to see
that Yahoo! searchmonkey lists 56.000 of them (search Yahoo for
"searchmonkey:com.yahoo.page.uf.recipe"). The site I work for is responsible
for about 12.000 of them but when i checked the first 70 results none of
them was from our site :-) In fact they almost all came from personal sites,
blogs and the like and this is really great! It was also very interesting to
see that a lot of them came from WordPress Blogs which shows that the
WordPress Plugin for hRecipe is very useful. But back to the issue: I
checked the first 60 or so results and about 2/3 of them where really using
hRecipe. 17 of them were from different sites(*) (**).

It's never bad to gather some empirical data to make an informed decision,
so I took the burden to count element use. Actually this didn't take much
more then an hour, so if somebody wants to check the results from 61 to ...
feel free! Well, now, without futher ado, the results (***):

 16   fn    
 17   ingredient (3 value, 3 type)
  3   yield 
 15   instructions 
  3   duration 
  4   photo
  9   summary
  4   author
  3   published
  1   nutrition (0 value, 0 type)
  2   tag

Of course this is not representative, but it gives a fairly good impression.
If you add our site (or 20% of all published hRecipes ;) you can add +1 to
all elements since we use everything...


Evaluation:

"fn", "ingredient" and "instructions" are no-brainers. Also "summary" is
surprisingly strong and the intuition, that it's a strong part of how people
communicate on recipes, seems to be right.

"author" and "published" probably didn't get so much used because on a blog
they are provided anyway and most of the sites investigated were blogs. OTOH
they aren't important for the functionality of a recipe and can be added by
other means. So maybe they should be removed for the sake of terseness.

"photo" was definitely less used than there were photos added to recipes.
And I know a lot of people who, when showed 2 recipes, one with and one
without photo, almost certainly choose the one with photo. So IMHO it should
stay.

"yield" was used surprisingly seldomly. Isn't it an essential addition to
the ingredients list? Maybe the yield is often self-explanatory, maybe this
reflects the fact, that also the value and type of ingredients wasn't marked
up most of the time. 'Smallishness' would certainly suggest to omit all the
three: "value", "type" and yield". People seem to feel that the "ingredient"
field naturally holds information not only about the name of the ingredient
but about type and value too and don't bother to mark these up explicitely.
Again for big sites which generate their content from databases the
cost-usage relation is totally different...

"duration" would be a candidate for removal as well. But still, remeber that
this list only reflects usage of the elements. An information of duration
was part of the recipe more often than it was marked up as such.

"tag" obviously didn't get used much and is maybe strong enough on it's own.
People add tags no matter if they are part of a vocabulary, and the rel-tag
pattern is certainly easy enough to grasp and straightforward to implement.
But the WordPress plugin added a tagging vocabulary on it's own (how dare
they... ;-) which did get used most of the time. So tagging principally is
strong. Well, I still don't have a clue on this one.

"nutrition" is definitely a candidate for removal, maybe into it's own
vocabulary. OTOH, big commercially backed sites very often provide
nutritional information and when provided it is a much more integral part of
the information context than say the author. So maybe it should stay?


Comments, thoughts?

Thomas 




(*) well, more of them, I skipped some WordPress Blogs since they all came
from the same Plugin. But as the feature set of the Plugin continues to grow
it doesn't make much sense to investigate them too thoroughly now unless you
want to take into account which version of the plugin was used exactly etc
etc - an undertaking I'm not really up to ;-)

(**) I felt quite bad when I saw that a lot of them used the pre-0.2 version
with "title", "method" etc but I really don't know what I could do about it
now (or then, too). I guess these decisions are made and the pain will
diminish over time (cough).

(***) The full result-set, including URLs, is on the
microformats.org/wiki/hrecipe-issues page.




.
Thomas Lörtsch
Business Development
G+J Exclusive&Living digital GmbH
Redaktion Online
..
Stubbenhuk 5
20459 Hamburg
...
eMail: loertsch.thomas at guj.de