Re: Enquire - WWW - Semantic Desktop/PersonalDataWiki? do you agree?

Hi Leo,
I seem to remember we had an exchange some time ago about Nepomuk/Semantic
Desktop and the complexity of developing file parsers. I think you were
having a glass of wine somewhere.
I am off work at the moment as I have just fractured my arm cycling (damn
it!).
I am very interested in semantic technologies, although I have no use for a
semantic desktop.
I am interested in the benefits and trade offs in using a semantic approach
to solve this problem:-
There are arbitrary in coming streams of data, a few although some are
sizeable data sets.
Some change rapidly (within seconds) others remain stable over months.
The formats are disparate, from XML (with or without schema) through CSV to
SQL calls into a db.
Our approach is to model the domain strongly, but at a sufficient level of
abstraction such that calls to the merged data set are uniform for the
client.
Then, for each source we use some form of ETL - depending on our needs a lot
of effort then has to go into each ETL.
My investigation into semantic solutions include -
1. The automacisity of conversion from XML to RDF.
2. Storage of RDF.
3. Remodelling RDF to reflect the domain.
4. Speed of query triple store against conventional DB.

1. I know of only 1. tool that achieves this - Topbraid - and there is
nothing wrong with this, but difficult to experiment as a skunk project
since it requires a licence.
I don't understand the degree of communality of approach in this area or how
difficult it would be for me to produce my own 'lifter'. Ultimately
desirable for control of the process.
2. This is related to 1. and 4. I'm thinking going the graph db way using
gremlin and any backend. I haven't found other approaches yet that promise
to tackle the issue of query response.
3. This should be the big gain. I am right off the extent of my
understanding but am just about to experiment with this today. I have done
1. and, for sensible queries, need 3. I believe it should be very straight
forward and should be the area in which this approach excels.
4. I have mentioned. This should be measured in a similar way to the Berlin
SPARQL benchmark.

My enquiry is in the direction of whether the very time consuming ETL phase
plus ORM development phase can be side stepped.
One of the issue is whether in coming data can be pre-processed more readily
and timely in this way.

I think this is a fair representation of some of the sorts of issues that
companies might turn to semantic technologies for.

What is important in all of this is that a company will take a proprietary
interest in the data they meld, change and serve.
I assume this is so with google and 'goodrelations'. They must see a benefit
to themselves in offering this format of data.

You have asked about the vision for a particular solution.
Are you lacking what is essential at this stage:a very clearly articulated
vision?
I don't have that vision. I see it in terms of data gathering and
enhancement along with a path to realise the benefits of that for the
company that adopts this solution.

There is, supposedly, a network effect and everyone wins now we have faxes.
That would be more like you can't afford not to have one. Not an assessment
of the actual benefit. And the semantic effort hasn't reached that point
yet. Certainly not for the desktop.
One needs a very clear idea of what problems one's solution tackles and what
the market area is in which these problems are found.
A semantic desktop seems to touch on other solution areas such as social
networking, open (or closed?) solutions in that area as well as machine
machine intelligent interaction.
I don't know how, if or the extent to which you are tackling those areas.
I do think there is a huge difference between enabling a user to create more
data in the form of links to learning about a user a la facebook or google.
I have to understand, what, where, is the value in the former?

I may be very old fashioned, but I see a public sphere increasingly
dominated by few large players principally interested in peoples' shopping
habits.
It requires little intelligence to choose between a Gucci bag or a pair of
Church's shoes or something.
It does require intelligence to do many of the professions, from solicitor
to surveyor. To help those people solve problems for their clients.
Is this the place for Gnowsis? If so how is it going to benefit in these
areas?

BTW, I run a mile from those who claim lack of vision, I see that as an
appeal to the most primitive herd instinct. So we all end up in one herd or
another - but so what!

Best,

Adam


On 18 November 2010 10:25, Leo Sauermann <leo.sauermann@gnowsis.com> wrote:

> Right, actually I am not dishearted, It is  a nice Job to pitch Gnowsis.
> Also, we have good feedback from customers who can now link emails to other
> things.
>
> Historically there is a chain of memex/xanadu/enquire/web.
> Part of the reason I wrote this mail is because I see a pattern here
> leading to personal semantic web. I am curious  whether this pattern is also
> seen by others.
>
> I don't plan to tell these things to investors, its rather an understanding
> I need for myself. So ... calling for the elders.
>
> Cheers from the business angel day austria. :-)
>
> Leo Sauermann, Dr.
> CEO and Founder
>
> mail: leo.sauermann@gnowsis.com
> mobile: +43 6991 gnowsis
> http://www.gnowsis.com
>
> helping people remember,
>
> so join our newsletter
> http://www.gnowsis.com/about/content/newsletter
>
> Am 18.11.2010 um 10:36 schrieb Nathan <nathan@webr3.org>:
>
> > Hi Leo!
> >
> > You sound disheartened, don't be :) Remember it's all about the link, the
> web is only a web because it's got links, twitter rose to greatness on the
> back of the typed link, you make one small link relation between you and
> another person, and look what it enables. Some have let people both read and
> write /and/ make links, like wikipedia and blogs, and look how that worked
> out! Other companies have software which make links between words and pages,
> like google, and look what that enables!
> >
> > And now you and Bernhard have brought the power of the link to peoples
> personal information on their desktop, with software that can make the links
> for you, look what that enables :)
> >
> > It's all right here:
> >  http://www.w3.org/2007/Talks/1211-whit-tbl/
> >
> > And the futures exciting:
> >  http://videolectures.net/www09_bernerslee_rlty/
> >
> > Especially when it's mapped out:
> >  http://www.w3.org/2010/roadmap/sem-seed.svg
> >
> > And when the path has been known for so long:
> >  http://www.w3.org/History/1989/proposal.html
> >
> > when the way is known:
> >  http://www.w3.org/DesignIssues/ReadWriteLinkedData.html
> >
> > and look what it enables:
> >  http://www.w3.org/DesignIssues/CloudStorage.html
> >
> > It's great to see what you're doing Leo, and Bernhard, and even better to
> know that you can see the web for what it is, enquire++.
> >
> > Keep up the good work :)
> >
> > Nathan
> >
> > Leo Sauermann wrote:
> >> Dear Tim,
> >> SWIG
> >> My Current Task is somehow tricky:
> >> I am pitching our company Gnowsis to investors. It makes an Enquire-Like
> >> Semantic Desktop, a personal semantic web. It seems investors
> >> only understand anything on the level of "we do twitter for dogs" or "it
> >> sells crowdsourced clothes via mobile phones".
> >> So Tim, as Elder of the Web, I turn to you for an expert opinion to
> >> reassure we do a useful thing.
> >> Here are the Statements I patch together from "weaving the web" etc.:
> >> Enquire - linking everything bidirectionally is an
> >> entire new way of writing.
> >> I guess you also realized that the system changes the way you look at
> >> things and your
> >> thinking.
> >> WWW - give everyone a tool (read/write) web that everyone
> >> can publish information. The links are first unidirectional and untyped
> >> and will be typed
> >> "later", once the RDF riddle was solved in 1999.
> >> SemanticDesktop/PersonalDataWikis/PersonalSemanticWeb - we finally give
> >> the peoples the
> >> Enquire that Tim already used in the distant and mysterious past.
> >> The first distributable results are NEPOMUK-KDE, PesonalDataWikis,
> >> Personal Data Lockers, and services around this idea of personal
> >> semantic web services for personal information management.
> >> The question:
> >> Tim,
> >> Is this THE idea?
> >> Do you agree this is a sensible thing to do?
> >> If yes, then my argumentation to investors and to myself is:
> >> "TimBl basically inventend blogging, wikipedia, and twitter with the
> >> idea of a read/write WWW, which it originally always included. You can
> >> trust that guy to be clever. You can also be assured this was around
> >> before, some ideas for millenia, some since Memex.
> >> Now Gnowsis works to realize the proto-idea - Enquire. There must be
> >> something going on here. Dear Investors, look at it, spend some time
> >> understanding what happens here with technology and then invest."
> >> Sometimes I feel like Frodo and together with Bernhard "Sam" Schandl we
> >> go alone to Mount Doom ("Microsoft Outlook") to finally throw the Ring
> >> of RDF into its center, to crack it open to the web. Then I see you guys
> >> over at the Minas Tirith of LinkedOpenData and data.gov and the battles
> >> Martin Hepp fights with GoodRelations and ... there is hope :-)
> >> ok, looking for interesting answers
> >> best
> >> Leo Sauermann, Dr.
> >> CEO and Founder
> >> mail: leo.sauermann@gnowsis.com
> >> mobile: +43 6991 gnowsis
> >> http://www.gnowsis.com
> >> helping people remember,
> >> so join our newsletter
> >> http://www.gnowsis.com/about/content/newsletter
> >> ____________________________________________________
> >
> >
>
>