FHIR Chat · Validation Pack downloads · implementers

Stream: implementers

Topic: Validation Pack downloads


view this post on Zulip Grahame Grieve (Jun 01 2016 at 06:01):

Currently, the validation pack downloads are split into 4 different downloads:

view this post on Zulip Grahame Grieve (Jun 01 2016 at 06:02):

validation.X.zip where X is XML orJSON

view this post on Zulip Grahame Grieve (Jun 01 2016 at 06:02):

and validation-min.X.zip, which are the same, but with all the descriptions and long text fields removed

view this post on Zulip Grahame Grieve (Jun 01 2016 at 06:03):

validation-min is provided because loading it takes considerably less memory.

view this post on Zulip Grahame Grieve (Jun 01 2016 at 06:03):

right now, the xml variant contains the conformance resources + W3C XML schema + schematron

view this post on Zulip Grahame Grieve (Jun 01 2016 at 06:04):

the json variant contains conformance resources + JSON schema

view this post on Zulip Grahame Grieve (Jun 01 2016 at 06:04):

but it turns out that the validator will only load the xml variant. Found this yesterday, when I get around to trying it with JSON - no one has said anything

view this post on Zulip Grahame Grieve (Jun 01 2016 at 06:05):

so do I conclude that no one uses the json variant?

view this post on Zulip Grahame Grieve (Jun 01 2016 at 06:05):

how should I package all this up, especially now that there's turtle/ShEx as well?

view this post on Zulip Jingdong(JD) Li (Jun 01 2016 at 07:42):

so far, only used xml validation, never tried json validation.

view this post on Zulip David Hay (Jun 01 2016 at 14:30):

I use the HTTP servers for validation (ie the $validate operation) because I'm in a browser world and happy to offload the work to someone else!

view this post on Zulip Lloyd McKenzie (Jun 01 2016 at 17:03):

Does the validator use the xsd or sch files anymore?

view this post on Zulip Grahame Grieve (Jun 01 2016 at 20:04):

no. They don't provide additional validation that the validator doesn't do

view this post on Zulip Lloyd McKenzie (Jun 01 2016 at 20:37):

Should we split them out too then?

view this post on Zulip Grahame Grieve (Jun 01 2016 at 22:48):

we already do

view this post on Zulip Lloyd McKenzie (Jun 03 2016 at 01:46):

Are they not in validator.zip anymore?

view this post on Zulip Grahame Grieve (Jun 03 2016 at 01:46):

yes they are right now

view this post on Zulip Grahame Grieve (Jun 03 2016 at 01:47):

I think I'm going to split it differently: just have definitions xml and json, and 3 schema downloads. no min.

view this post on Zulip Lloyd McKenzie (Jun 03 2016 at 01:54):

I think having a min is still useful

view this post on Zulip Grahame Grieve (Jun 03 2016 at 01:55):

why?

view this post on Zulip Lloyd McKenzie (Jun 03 2016 at 01:56):

Because at runtime, you don't want all of the text. The lighter we make validation, the more people can do it.

view this post on Zulip Lloyd McKenzie (Jun 03 2016 at 01:57):

To be honest, you might get away with only having the "min"

view this post on Zulip Grahame Grieve (Jun 03 2016 at 01:57):

who's doing validation? The only users of validation use the jar

view this post on Zulip Grahame Grieve (Jun 03 2016 at 01:57):

the jar will include it's own copy, minified

view this post on Zulip Lloyd McKenzie (Jun 03 2016 at 02:01):

I think it's useful to have the code and instances separate so that you can in theory run the same code against different versions of the resource. But that is an "in theory" thing.

view this post on Zulip Grahame Grieve (Jun 03 2016 at 02:01):

why would that be useful? if you're capable of modifying the validator code, you're capable of managing the packaging issue, or working with the full definitions

view this post on Zulip Erich Schulz (Jun 03 2016 at 07:37):

i am currently developing a javascript app I'm hoping to run on phones @Grahame Grieve - I imagine some of the metadata would be useful for validation in that contect - but possibly only a subset - if the entire validation pack comes in small enough, is this a usecase?

view this post on Zulip Erich Schulz (Jun 03 2016 at 08:12):

but @Grahame Grieve I agree it maybe unnecessary to distribute both versions as the I'd expect the "minification process" to be only 10 lines of code...

view this post on Zulip Grahame Grieve (Jun 03 2016 at 10:33):

bit more than 10 lines to express the logic, but I don't think it's realistic to perform validation on the phone. Better just to pass it to a server - quicker, too, because the server can be a full blown terminology server too

view this post on Zulip Erich Schulz (Jun 03 2016 at 10:41):

cool - this is the luxury of JS - I wont have to really lock that in...

view this post on Zulip Erich Schulz (Jun 03 2016 at 10:43):

but if the "normal" to "min" transformation is just "the same, but with all the descriptions and long text fields removed" then that is the complexity from a large number of inconsistently named long fields?

view this post on Zulip Erich Schulz (Jun 03 2016 at 10:44):

recursively walking a tree and stripping all fields called "notes" and "description" is just two lines...

view this post on Zulip Grahame Grieve (Jun 03 2016 at 10:49):

requirements, copyright, a few others

view this post on Zulip Pascal Pfiffner (Jun 15 2016 at 15:52):

I use validation-min.json.zip in the parser that creates the Swift and Python classes. I've been using profiles-types.json and profiles-resources.json from that pack. Is that available elsewhere?

view this post on Zulip Grahame Grieve (Jun 15 2016 at 19:36):

you mean are those files available elsewhere? well, I've renamed the download 'definitions.json.zip', but you can also access them directly

view this post on Zulip Pascal Pfiffner (Jun 16 2016 at 13:55):

And I see they are also in examples-json.zip, which I download to run unit tests, so I'm all good!


Last updated: Apr 12 2022 at 19:14 UTC