Collection web service (TAPIR, DiGIR) status
14 views
Skip to first unread message
Rod Page
Sep 18, 2008, 6:32:21 AM9/18/08
to Biodiversity Collections Index
I guess this is a feature request. Given that the Biodiversity
Collections Index will potentially list all DiGIR/TAPIR providers, how
about showing the current status of those providers?
Dave Vieglais' BigDig http://bigdig.ecoforge.net/ did this for a
while, but doesn't seem to have been updated since February this year.
I wonder if having a symbol beside the web service would prompt the
people responsible for these services to ensure they are online (base
don the assumption that BioCol is a visible resource). I'm constantly
amazed by how many are offline (or change web server, or schema
version, or resource code). One could even use a sparkline to show
status history over time. The status checking itself doesn't need to
be too fancy. Whether the service returns a 404 would be a start,
followed by more sophisticated error checking if need be. Dave may
have some code, or I could write something.
What prompted me to suggest this is I just edited the Smithsonian fish
collection record, and the DiGIR provider is
http://nhb-acsmith2.si.edu/emuwebvzfishesweb/webservices/digir.php .
Having a web server called nhb-acsmith2.si.edu is asking for it to
break (OK, maybe I'm being unfair, there's a redirect to there from
http://acsmith.si.edu/emuwebvzfishesweb/webservices/digir.php, but you
get my point).
Collections Index will potentially list all DiGIR/TAPIR providers, how
about showing the current status of those providers?
Dave Vieglais' BigDig http://bigdig.ecoforge.net/ did this for a
while, but doesn't seem to have been updated since February this year.
I wonder if having a symbol beside the web service would prompt the
people responsible for these services to ensure they are online (base
don the assumption that BioCol is a visible resource). I'm constantly
amazed by how many are offline (or change web server, or schema
version, or resource code). One could even use a sparkline to show
status history over time. The status checking itself doesn't need to
be too fancy. Whether the service returns a 404 would be a start,
followed by more sophisticated error checking if need be. Dave may
have some code, or I could write something.
What prompted me to suggest this is I just edited the Smithsonian fish
collection record, and the DiGIR provider is
http://nhb-acsmith2.si.edu/emuwebvzfishesweb/webservices/digir.php .
Having a web server called nhb-acsmith2.si.edu is asking for it to
break (OK, maybe I'm being unfair, there's a redirect to there from
http://acsmith.si.edu/emuwebvzfishesweb/webservices/digir.php, but you
get my point).
rogerhyam
Sep 18, 2008, 6:54:08 AM9/18/08
to Biodiversity Collections Index
Hi Rod,
Good to have you on board.
Yes I think this is a good possibility for future enhancement. I see
two main issues - neither technical.
1) There is some odd non 1:1 relationship between what is considered a
collection and a DiGIR/TAPIR end point. Looking through lists of DiGIR
providers they seem to overlap collections or represent parts of
collections. If we create collections in BCI that just exist to
represent the coverage of a DiGIR provider then we are kind of
becoming a service registry rather than a list of real world things -
which worries me slightly.
2) The GBIF data portal probably already knows bout all the DiGIR/
TAPIR/BioCASe providers out there and is probably scraping them
already and knows when they are down so that might be the natural
place to put a list of providers and their current status rather than
building anything new.
If we get the web service fields filled in for a reasonable number of
collections I would, at a minimum, write a web service to retrieve a
list of them so that any third party could write their own polling
mechanism. In fact if BCI were to take on such a role that is probably
how I would implement it anyway. I'll bear this in mind if I ever find
time to expand the JSON service - soon I hope.
Does that make sense?
Roger
On Sep 18, 11:32 am, Rod Page <r.p...@bio.gla.ac.uk> wrote:
> I guess this is a feature request. Given that the Biodiversity
> Collections Index will potentially list all DiGIR/TAPIR providers, how
> about showing the current status of those providers?
>
> Dave Vieglais' BigDighttp://bigdig.ecoforge.net/did this for a
> get my point).
Good to have you on board.
Yes I think this is a good possibility for future enhancement. I see
two main issues - neither technical.
1) There is some odd non 1:1 relationship between what is considered a
collection and a DiGIR/TAPIR end point. Looking through lists of DiGIR
providers they seem to overlap collections or represent parts of
collections. If we create collections in BCI that just exist to
represent the coverage of a DiGIR provider then we are kind of
becoming a service registry rather than a list of real world things -
which worries me slightly.
2) The GBIF data portal probably already knows bout all the DiGIR/
TAPIR/BioCASe providers out there and is probably scraping them
already and knows when they are down so that might be the natural
place to put a list of providers and their current status rather than
building anything new.
If we get the web service fields filled in for a reasonable number of
collections I would, at a minimum, write a web service to retrieve a
list of them so that any third party could write their own polling
mechanism. In fact if BCI were to take on such a role that is probably
how I would implement it anyway. I'll bear this in mind if I ever find
time to expand the JSON service - soon I hope.
Does that make sense?
Roger
On Sep 18, 11:32 am, Rod Page <r.p...@bio.gla.ac.uk> wrote:
> I guess this is a feature request. Given that the Biodiversity
> Collections Index will potentially list all DiGIR/TAPIR providers, how
> about showing the current status of those providers?
>
> while, but doesn't seem to have been updated since February this year.
> I wonder if having a symbol beside the web service would prompt the
> people responsible for these services to ensure they are online (base
> don the assumption that BioCol is a visible resource). I'm constantly
> amazed by how many are offline (or change web server, or schema
> version, or resource code). One could even use a sparkline to show
> status history over time. The status checking itself doesn't need to
> be too fancy. Whether the service returns a 404 would be a start,
> followed by more sophisticated error checking if need be. Dave may
> have some code, or I could write something.
>
> What prompted me to suggest this is I just edited the Smithsonian fish
> collection record, and the DiGIR provider ishttp://nhb-acsmith2.si.edu/emuwebvzfishesweb/webservices/digir.php.
> Having a web server called nhb-acsmith2.si.edu is asking for it to
> break (OK, maybe I'm being unfair, there's a redirect to there fromhttp://acsmith.si.edu/emuwebvzfishesweb/webservices/digir.php, but you
> I wonder if having a symbol beside the web service would prompt the
> people responsible for these services to ensure they are online (base
> don the assumption that BioCol is a visible resource). I'm constantly
> amazed by how many are offline (or change web server, or schema
> version, or resource code). One could even use a sparkline to show
> status history over time. The status checking itself doesn't need to
> be too fancy. Whether the service returns a 404 would be a start,
> followed by more sophisticated error checking if need be. Dave may
> have some code, or I could write something.
>
> What prompted me to suggest this is I just edited the Smithsonian fish
> collection record, and the DiGIR provider ishttp://nhb-acsmith2.si.edu/emuwebvzfishesweb/webservices/digir.php.
> Having a web server called nhb-acsmith2.si.edu is asking for it to
> get my point).
Rod Page
Sep 18, 2008, 7:12:01 AM9/18/08
to Biodiversity Collections Index
Roger,
All good points, and I know it's important to resist feature creep.
GBIF internally will know which services are online, but I've not seen
an obvious way to find out from their site. I agree that the mapping
may be problematic, but from my own perspective the digital collection
is what matters most. Physical collections are great, but without
digital access they are essentially invisible. I may add some more
DiGIR endpoints to the database as I come across them.
Rod
> > Dave Vieglais' BigDighttp://bigdig.ecoforge.net/didthis for a
All good points, and I know it's important to resist feature creep.
GBIF internally will know which services are online, but I've not seen
an obvious way to find out from their site. I agree that the mapping
may be problematic, but from my own perspective the digital collection
is what matters most. Physical collections are great, but without
digital access they are essentially invisible. I may add some more
DiGIR endpoints to the database as I come across them.
Rod
Mike Giddens
Sep 19, 2008, 10:03:54 AM9/19/08
to Biodiversity Collections Index
The easiest way for the Tapir checking (since you use php/zend on that
server) could be to use TapirChirp to just ping the service or just
write a simple ping request. I also agree that some of the points of
concern are valid. I do think it would be nice to know where the
tapir/digir service is for a collection but it might be easier to work
with Gbif to crossreference the uddi domain registration to the
domains you have registered. This might load in a good base.
-Mike
> > > Dave Vieglais' BigDighttp://bigdig.ecoforge.net/didthisfor a
server) could be to use TapirChirp to just ping the service or just
write a simple ping request. I also agree that some of the points of
concern are valid. I do think it would be nice to know where the
tapir/digir service is for a collection but it might be easier to work
with Gbif to crossreference the uddi domain registration to the
domains you have registered. This might load in a good base.
-Mike
Reply all
Reply to author
Forward
0 new messages