Re: PiggyBank install feedback

From: David Karger <karger_at_mit.edu>
Date: Sun, 30 Oct 2005 09:49:58 -0500

David Huynh wrote:

> David Karger wrote:
>
>> Michael McDougall wrote:
>>
>>> Stefano Mazzocchi wrote:
>>>
>>>> Michael McDougall wrote:
>>>>
>>>>>
>>>>> - The metaphor seems a little discontinuous to me. When I click
>>>>> the "RDF coin" the animation (nice, BTW) shows RDF coins going
>>>>> into a slot--it looks like the RDF (coin) is being stored in my DB
>>>>> (piggy bank). But then on the next page I find out this isn't the
>>>>> case--I still need to save things before the data is *really*
>>>>> stored in my piggy bank. I could see how people may assume the
>>>>> data is stored in the bank already and "Save" means save the item
>>>>> to a file.
>>>>
>>>>
>> It's a bit strange actually. When I view a web page in my browser,
>> do I have sense that the page is "saved" as a result? Why should
>> viewing the "rdf-version" of a web page be any different? Perhaps
>> there should be no animation at all when I "change perspective" and
>> instead the coin-in-slot should happen during the save..
>
>
> Well, for Web pages, we do have bookmarking and saving locally (in
> Firefox, File -> Save Page As).

We don't really have an analogue to bookmarking in pb (namely, when I
"return" to a resource in my pb, I see the rdf I saved in my pb---NOT
the possibly changed rdf currently associated with that resource at
whatever site I visited to get it in the first place). So it is clear
that PBs model is more of a "save page as". However, that is relatively
rarely used and certainly is not what people thinks happen when they
just visit a page.

>
> Perhaps this is what should happen in future version of Piggy Bank:
>
> - If a Web page voluntarily offers RDF (as RDF/A, GRDDL, or through
> some conventional <link>), then Piggy Bank automatically retrieves
> that data and augments the page with semantic elements (e.g.,
> semantic-sensitive context menus and such). A Web page usually does
> not contain too much data, and so this process of retrieval and
> augmentation hopefully does not negatively affect the user's browsing
> experience. Good things just happen. The user doesn't even have to
> switch view just yet.
>
> - If a Web page does not offer RDF, then Piggy Bank indicates whether
> there is any applicable scraper but does not automatically proceed to
> scrape. The user has to invoke the scraping operation, which will take
> some time.
>
I think this is a false distinction. If the amount of rdf is small,
then whether one gets direct or scrapes, it will be fast. if it is
large, either way will be slow.

Actually, this suggests a good heuristic for pb: scrape with a time
limit. If you finish in time, great. If not, give up until user
explicitly requests a scrape.

and anyway, this is all clearly coping with an
implementation/performance detail, and I dislike the idea of having the
user interface twisted out of shape by such a temporary issue---because
changing the ui after we fix the performance issue will confuse people.
Long term, we should just fetch rdf in the background in all cases, right?

> - If there is no RDF and no scraper, (or if the user so wants), the
> user can pull out Piggy Bank's do-it-yourself scraping UI and perform
> the just-in-time scraping.

"just in time" implies it happens without asking. once you have to ask,
it is too late :)

>
> These 3 levels of integration hopefully will make clear the
> experiential distinction between a SW-friendly Web site and a
> SW-unfriendly Web site. Users are hopefully driven over to SW-friendly
> sites, making demand on more sites to offer RDF.
>
>
> Saving into "My Piggy Bank" is then a separate issue altogether.
> Saving will be the combo of bookmarking and Save Page As. It will be
> used for both remembering the things you have


As mentioned above, "save" and "bookmark" are fundamentally different:
one freezes the data while the other holds a pointer to a dynamic
version. We shouldn't blend these two. Granted that it is not clear
exactly what bookmarking means for rdf, it certainly isn't the same as
saving.

> encountered earlier so to have a convenient way to quickly get back to
> it, and for capturing the current state of that thing lest it might be
> changed by the original site in the future.
>
> David
>
>>>> Hmmm, good point actually. Hmmm maybe we could have the coin fall
>>>> from the sky and pile up instead and having the "save all content
>>>> in this page" (see below) use the 'coins in the slot' one. Warning,
>>>> I think we are approaching very fast my abilities as an icon
>>>> designer btw :-)
>>>
>>>
>>> How about this?
>>> When you collect you could show the coins going into a pocket.
>>> When you save you could show the coins going from your pocket into
>>> the piggy bank.
>>>
>>> The metaphor's not perfect: coins in my pocket don't disappear if I
>>> fail to put them in my piggy bank right away. But it does kind of
>>> fit the way I accumulate coins.
>>>
>>> Michael
>>
>>
Received on Sun Oct 30 2005 - 14:44:41 EST