From - Tue Jan 28 15:42:07 2003 X-UIDL: 3e36a2bd00000001 X-Mozilla-Status: 0013 X-Mozilla-Status2: 00000000 Received: from pantheon-po03.its.yale.edu (pantheon-po03.its.yale.edu [130.132.50.19]) by ns.cria.org.br (8.11.2/8.11.2) with ESMTP id h0SFWgi11446 for ; Tue, 28 Jan 2003 13:32:43 -0200 Received: from pspig.its.yale.edu (pspig.its.yale.edu [130.132.50.12]) by pantheon-po03.its.yale.edu (8.11.6/8.11.6) with ESMTP id h0SGWbn01312; Tue, 28 Jan 2003 11:32:37 -0500 (EST) Received: (from imphttpd@localhost) by pspig.its.yale.edu (8.11.6/8.11.6) id h0SGWas28511; Tue, 28 Jan 2003 11:32:36 -0500 (EST) Received: from 130.132.27.100 ( [130.132.27.100]) as user rb377@rb377.mail.yale.edu by www.mail.yale.edu with HTTP; Tue, 28 Jan 2003 11:32:36 -0500 Message-ID: <1043771556.3e36b0a4440a1@www.mail.yale.edu> Date: Tue, 28 Jan 2003 11:32:36 -0500 From: reed.beaman@yale.edu To: Ricardo Scachetti Pereira Cc: "Stewart, Aimee Marian" , "Beach, James H" , "Vorontsov, Gregory Yakovlevich" , "Downie, Scott Russell" , reed.beaman@yale.edu Subject: Re: Value added products to collection as webservices References: <3E352105.2090105@cria.org.br> In-Reply-To: <3E352105.2090105@cria.org.br> MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit User-Agent: Internet Messaging Program (IMP) 3.1 X-Originating-IP: 130.132.27.100 X-YaleITSMailFilter: Version 1.0c (attachment(s) not renamed) X-MailScanner: Found to be clean Status: O Nice work Ricardo! On Item 2 I'm wondering whether proper fielding of these items (e.g., getting country, state, province, adm2, etc into the right fields) is covered here, or should it be a separate service. I've noticed this to be a more serious problem than I originally thought it to be, so much so that a georeferencing service has to cross-check fields to work properly (geomancer doesn't do this, yet). Greg I believe has noticed that there are lots of records with state/province in the country field and vice-versa. Reed Quoting Ricardo Scachetti Pereira : > It is good that you liked, Aimee. > See answers below: > > Stewart, Aimee Marian wrote: > > > Hi Ricardo, > > > > Great suggestions! I think that this is an excellent way to > increase > > our value to data providers. I will think about some more > services > > that can be implemented. > > > > By the way, 2 services are available > > at http:\\beta.lifemapper.org\Services\QueryByLocation > > and http:\\beta.lifemapper.org\Services\QueryTaxa . Number 6 on > your > > spreadsheet could be easily be put together from methods in these > > > services. > > You are right! > There is also another implementation of this service that creates > many > models, jackknifing the input datapoints (submitted paper by Town and > > I), that might be interesting, too. Reed is involved on this one, > too. > > > > > Number 7 would be easy once we gather layers for areas outside the > > > US. I will get to that once we have successfully moved to > production. > > Yeah! The data repositories that those kind of services rely on can > be > build in phases, as more data is added on. For the parts of the world > > that we don't have data, we output a diagnostic message: no data > available. Those datasets can also be used to normalize those same > data > fields. > > > > > We can create the initial services, then create some applications > > > using the services to implement commiting the new data back to > > provider databases (interactively, one record at a time or groups > at a > > time for normalization, in Specify perhaps). To get some interest > > > quickly, I think we need to allow providers a way to improve their > own > > data without on-site programmers. > > Right. > > > > > Report cards could include geostatistical analysis of entire > > collections, genus, species, subspecies. I've been wanting to > explore > > that ESRI extension! > > Yeah! I've been thinking about measuring correlation between a given > set > of points (per species, genus, museum, collector, etc) and features > > related to infrastructure (roads, rivers), and also with county > centroids. That would give a basis for comparing sampling between > datasets. Also, we could measure correlation of datapoints and > ecological dimensions, but that is for other purposes. > In this line, I was also thinking about some kind of density measures > > (spatial and temporal, such as, points per square mile, or per year), > > and also measures of clustering. > > > > > Perhaps a short questionaire for existing contributors could > determine > > whether there is any interest in these services, or if they have > some > > ideas we haven't yet come up with. The initial Fishnet (Ed) > reaction > > was very positive! > > Did you show that to Ed? Cool!!! > > > > > Aimee > > > > > > > > -----Original Message----- > > From: Ricardo Scachetti Pereira [mailto:ricardo@cria.org.br] > > Sent: Friday, January 24, 2003 11:47 AM > > To: Beach, James H; voron999@UKANS.EDU; Stewart, Aimee Marian; > Downie, > > Scott Russell; reed.beaman@yale.edu > > Subject: Value added products to collection as webservices > > > > Hey, guys. > > How have you been? > > > > Let me bring something up, so we can discuss it and think > > about it a little, if you want. > > I've been thinking about value added products to > collections > > and came up with some ideas. > > Among the categories of value added products, I could list > those: > > 1) Error checking and correction tools, such as: > normalization > > of field values, checking for invalid characters, checking > > compliance of date or coordinate formats, checking mispellings > and > > typos, cross-reference of textual location descriptions > (country, > > state, county, etc) with georeferences, etc; Reed's > georeferencing > > tool could be put in this category, if you consider that > putting a > > georeference in a record that does not contain one means > > correcting it, i.e., filling in the blanks on georeference > field. > > 2) Visualization tools: Showing data on maps (like GIS > tools > > and WMS services); > > 3) Statistics and other analysis tools: Measuring > clustering > > of points, correlation between data points and > infrastructure, > > sampling densities, etc. > > As far as I know, the tools in category #3 don't lend > > themselves well to a webservice implementation. They could be > > implemented as services, but they are not very important to > stand > > up for themselves. > > Tools in #2 are currently being well covered by Aimee's > WMS. > > So I tried to think about possible #1 tools that would > become > > good webservices. The result of that is summarized in the > Excel > > spreadsheet attached. > > I'm still not pretty sure what would be the best > architecture > > to implement them. > > A possibility would be implement each of them as one > object, > > or methods within an object. > > Each object would have a SOAP and/or XML/HTTP wrapper so > they > > become webservices. > > This way, local server components can use the objects > > directly. Remote components and other users can call the > services > > programmatically using the webservices. Some can be implemented > in > > Kansas, other at CRIA, or we could have copies in both > places. > > Pretty flexible. > > Last but not least, there are the report cards to the > > collections, that could use some of those services and some > other > > statistics from Lifemapper. > > Questions: > > 1) What do you think of the general idea and the proposed > > architecture? > > 2) What do you think of the services listed? Can you add > some > > more to the list? > > 3) Do you guys have more ideas for the report cards? What > > should be in there? > > 4) Market analysis: would any one out there (collections > > community) in need to use such services? > > > > That is about it. > > Cheers, > > > > Ricardo > > > > > > > > Attached is a set of 10 possible value added products that > > > >-- > >Ricardo Scachetti Pereira > >Gerente de Pesquisa > >Centro de Referência em Informação Ambiental - CRIA > >http://www.cria.org.br/ > > > > > > > > -- > > This message has been scanned for viruses and > > dangerous content and is believed to be clean. > > > > > > -- > > This message has been scanned for viruses and > > dangerous content and is believed to be clean. > > > -- > Ricardo Scachetti Pereira > Gerente de Pesquisa > Centro de Referência em Informação Ambiental - CRIA > http://www.cria.org.br/ > > > > -- > This message has been scanned for viruses and > dangerous content and is believed to be clean. > > -- This message has been scanned for viruses and dangerous content and is believed to be clean.