Time |
S |
Nick |
Message |
14:33 |
|
pdurbin |
got a nice walk through of the dvn by our head QA person |
14:33 |
|
pdurbin |
one thing we looked at was data visualization. you can see a nice example from http://dvn.iq.harvard.edu/dvn/dv/recessiontrends |
14:36 |
|
pdurbin |
at this time data visualizations are only supported with time series data |
14:36 |
|
pdurbin |
which is a good start |
14:37 |
|
pdurbin |
it's a little... involved... for the owner of the study to set up their data to support visualization. but once it's set up it's quite easy for users to come along and make visualizations out of the data |
14:39 |
|
pdurbin |
here's how the graphs look: http://guides.thedata.org/book/explore-data |
14:40 |
|
pdurbin |
and here's a walk through of how the study owner would set up their time series data to be visualized: http://guides.thedata.org/book/set |
14:40 |
|
pdurbin |
from http://guides.thedata.org/book/data-visualization |
14:41 |
|
pdurbin |
also, we talked a bit about graphml data: http://en.wikipedia.org/wiki/GraphML |
14:42 |
|
pdurbin |
i knew the dvn supports graphml but i didn't have an example dataverse handy. now i do: http://dvn.iq.harvard.edu/dvn/dv/patent |
14:42 |
|
pdurbin |
this seems to be our main doc about graphml: http://guides.thedata.org/book/network-data |
14:46 |
|
pdurbin |
another feature we went through that i wasn't aware of was study templates: http://guides.thedata.org/book/manage-study-templates |
14:49 |
|
pdurbin |
this got us talking about DDI (Data Documentation Initiative), which has an XML specification at http://www.ddialliance.org/Specification/ |
14:49 |
|
pdurbin |
"The Data Documentation Initiative (DDI) is an effort to create an international standard for describing data from the social, behavioral, and economic sciences. Expressed in XML, the DDI metadata specification now supports the entire research data life cycle. DDI metadata accompanies and enables data conceptualization, collection, processing, distribution, discovery, analysis, repurposing, and archiving." -- http://www.ddialliance.org/what |
14:50 |
|
pdurbin |
that's a pretty good description. and there's a wikipedia page on it: http://en.wikipedia.org/wiki/Data_Documentation_Initiative |
14:54 |
|
pdurbin |
interesting... "The DVN imports and exports its study metadata in various XML formats — Data Descriptive Initiative (DDI, version 2), Dublin Core, FGDC, and MARC." -- http://www.dlib.org/dlib/january11/crosas/01crosas.html |
14:59 |
|
pdurbin |
"You can also add your own custom fields to the Data Collection/Methodology section of the template." -- the "Manage Network Study Templates" section of http://guides.thedata.org/book/network-options ... that was useful too... to see how to add custom fields |
15:40 |
|
pdurbin |
oh, it was good to hear how harvesting works too... that the metadata is carried over but the files, the original studies, still reside on the original source |
15:43 |
|
pdurbin |
so when i go to http://dvn.iq.harvard.edu/dvn/dv/odum for example it says "Files for this study may also be accessed from their original source" when i go to download a file. since http://dvn.iq.harvard.edu/dvn/ hosts the metadata for odum but not the files themselves |
17:12 |
|
pdurbin |
"check out new Databridge website - #NSF funded collaboration with @reachnchub @uncsils @AggieResearch & @IQSS http://databridge.web.unc.edu " -- https://twitter.com/reachnchub/status/288662085192273920 |
17:15 |
|
pdurbin |
i had mentioned DataBrige at http://irclog.iq.harvard.edu/dvn/2012-12-17#i_169 |
18:57 |
|
|
esotiri joined #dvn |
18:57 |
|
pdurbin |
esotiri: hi! |
19:00 |
|
esotiri |
Phil, dvn-vagrant looks great. thanks |
19:00 |
|
pdurbin |
esotiri: yeah? it actually worked?!? ;) |
19:00 |
|
esotiri |
notice: Finished catalog run in 905.65 seconds |
19:00 |
|
pdurbin |
\o/ |
19:01 |
|
* pdurbin |
looks at https://github.com/pdurbin/dvn-vagrant |
19:01 |
|
pdurbin |
esotiri: you can log in at http://localhost:8081/dvn ? |
19:04 |
|
esotiri |
yes! I'll see to add fields to the study template loged as networkadmin |
19:04 |
|
pdurbin |
nice |
19:05 |
|
pdurbin |
i love it when stuff works :) |
19:05 |
|
pdurbin |
and that demo is a hack. needs some clean up |
19:05 |
|
pdurbin |
"experimental" we keep calling it on the mailing list :) |
19:05 |
|
esotiri |
thank you! |
19:17 |
|
pdurbin |
sure :) |
21:59 |
|
|
simmel joined #dvn |
22:00 |
|
simmel |
pdurbin: Hit me up with questions and I'll answer them. testshib is def the way to go for testing. |
22:07 |
|
pdurbin |
simmel: yay! thanks |
22:09 |
|
pdurbin |
i'm still learning about shibboleth... to me it's primarily a way to *not* have to store passwords locally. does that make sense? like when i log into stackoverflow, i can log in with openid. or i can log into some services with twitter's oath |
22:11 |
|
pdurbin |
in our case, for the dvn, we store usernames and passwords in a postgres database |
22:11 |
|
pdurbin |
our main installation of dvn is here: http://dvn.iq.harvard.edu/dvn/ |
22:12 |
|
pdurbin |
if you click "login" you'll see "dataverse account" and "affiliate" |
22:12 |
|
pdurbin |
that "affiliate" link shows Harvard and MIT right now |
22:12 |
|
pdurbin |
"affiliate" has kind of a special meaning in our app... |
22:13 |
|
pdurbin |
... but from a shibboleth perspective, i wonder how the login page would look... |
22:13 |
|
pdurbin |
would there be a long list of institutions that are able to log in via shibboleth? |
22:15 |
|
simmel |
pdurbin: Depends if you are an SP or an IDP. But if you'e a SP then yes it's like being a site which uses openid. |
22:15 |
|
pdurbin |
so our app will be an SP |
22:16 |
|
pdurbin |
as an app developer i don't think i need to worry about being an IdP |
22:17 |
|
pdurbin |
here at harvard we aren't using shibboleth yet but other institutions that use our app are already using shibboleth |
22:17 |
|
pdurbin |
so i guess once we put this feature in... they could stop storing their user accounts in their local postgres db and somehow point the app at their shibboleth instead |
22:18 |
|
simmel |
With shibboleth you just use the scoped userid (real or a generated) to attach to a session or a "user" in your database. No need to store a password. |
22:18 |
|
pdurbin |
hmm, ok. makes sense that we would still have some information about the users in our database. app-specific stuff like what their role is in the app (i.e. if they are an admin or not) |
22:19 |
|
pdurbin |
it reminds me a little of ldap... often you can point an app at your ldap server and let people use their ldap password to log in |
22:20 |
|
simmel |
In Shibboleth you can "require" authentication on a specific UR[LI] or you can use "lazy sessions". When a user hit the part of your page which requires a session they get redirected either to harvards IDP to login (if you only want harvard users to be able to login). |
22:21 |
|
simmel |
The user can also get redirected to a WAYF/DS where the user can choose from a number of IDP's e.g. all .us unis which are available in InCommon |
22:22 |
|
pdurbin |
interesting |
22:22 |
|
simmel |
Often there is just one IDP per university since all persons have a university account (I guess? This can be different per uni or country) |
22:22 |
|
pdurbin |
maybe we would make that configurable... installers of our app use it in different ways |
22:22 |
|
simmel |
Oh, ok. |
22:23 |
|
simmel |
But doesn't harvard have an IDP of their own? |
22:23 |
|
pdurbin |
no. not yet |
22:23 |
|
simmel |
Oh, ok. |
22:23 |
|
pdurbin |
but |
22:25 |
|
pdurbin |
if i put in this feature, even in our own installation at harvard... we could let institutions that do use shibboleth to log in... right? |
22:25 |
|
pdurbin |
i'm already talking to unc.edu. they use shibboleth: http://its.unc.edu/service-catalog/shibboleth/ |
22:26 |
|
pdurbin |
anyway, i'm sorry but i have to run to pick up my kids. thanks for stopping by! i hope we can pick this up later |
22:27 |
|
simmel |
Read up on https://wiki.shibboleth.net/confluence/display/SHIB2/UnderstandingShibboleth everything under "Shibboleth Basics" and I think you will get the big picture |
22:28 |
|
simmel |
Also, on that page read everything under "Shibboleth SP" |
22:32 |
|
simmel |
pdurbin: Yes you could let institutions which use shibboleth to login. Do you know if there is a harvard WAYF/Discovery Service in place? Or if InCommon provides one for you? |
22:33 |
|
simmel |
pdurbin: Sure, just keep hitting me with questions and I'll answer them when I have time. |