HMIS 2.8 Draft for your review

[Eric Jahn]

Here is a link to the HMIS 2.8 XSD as it now stands (draft):
http://tinyurl.com/ypxjh8

A valid instance doc example: http://tinyurl.com/2hpddx

An example of another schema importing HMIS 2.8 and extending it:
http://tinyurl.com/ynjh46

An instance of the extending schema: http://tinyurl.com/2cy2vm

Wiki page with a discussion of planned changes:
http://tinyurl.com/2sxrd4

Looking forward to a good discussion of these modifications in this
channel. Thanks!

--
Eric Jahn
Alexandria Consulting LLC
3126 8th Ave. N
St. Petersburg, FL 33713
941.321.1466

[David Talbot]

First, I applaud starting to create SOAP based standards around
transmitting, receiving, error handling. Second, around enhancing the
extensibility can you give me a quick overview of what changes you're
looking to make for WI?

I took a quick scan of the 2.8 schema (skipping to the custom tag where I
expected to see changes) and couldn't spot what's different vs 2.7.

Also, you might change the tinyurl link to go to http instead of https (not
like this stuff is secret anyhow). I'm sure some folks heed the grave
warnings both IE and firefox give about a site's certificate being
incorrect. :)

[Eric Jahn]

On Tue, 2007-07-31 at 12:43 -0500, David Talbot wrote:
> First, I applaud starting to create SOAP based standards around
> transmitting, receiving, error handling.

David, thanks as always for your prompt feedback! Yes, among other
myriad problems with not having messaging standard, it's causing a lot
of difficulties by not having a convenient way to say "this is an
update" or "this is a deletion" record set.

> Second, around enhancing the
> extensibility can you give me a quick overview of what changes you're
> looking to make for WI?

David, by WI, you mean the Minnesota work? MN is using the approved
2.7, so some of the extensibility changes are a reflection of some of
the problems we encountered extending the HUD HMIS schema to create a
Minnesota extended schema. The big change in the 2.8 schema is that all
the elements are based on top level complex types. So everything is
importable into another schema. The extension example I linked to is
not possible with the existing 2.7 schema. Also, everything is very
clearly namespaced in 2.8, whereas 2.7 had no namespace. This should
help keep things visually (and actually) separated when importing.

> I took a quick scan of the 2.8 schema (skipping to the custom tag where I
> expected to see changes) and couldn't spot what's different vs 2.7.

The custom tag is likely to be dropped, and I'm glad you mentioned this.
Does anyone plan on using the custom tags moving forward? I think
importing/extending into a custom schema is the way to go instead, since
the resulting custom schema can be validated. The custom tag using
"any" is not able to be validated, and allows free-form structures in
instance XML docs.

Another big change is granular timestamping (as opposed to the
broad-brushed ClientHistorical timestamping). As far as all the
changes, the color-coded line-by-line revision history from 2.7 is
located at: http://tinyurl.com/372nhz

> Also, you might change the tinyurl link to go to http instead of https (not
> like this stuff is secret anyhow). I'm sure some folks heed the grave
> warnings both IE and firefox give about a site's certificate being
> incorrect. :)

Yeah, I need to tweak some server configs... Until I do, if anyone has
problems logging in, let me know. Thanks -Eric

[David King]

I haven't had a chance to review it entirely yet, but it looks looks
like a good step forward. I particularly like the dataStampGroup and
identifierStampGroup applied to most Client and clientHistorical
elements. Not sure why you left identifierStampGroup off Gender?

I also like the restructuring so that it can more easily be modfied.
A much better replacement to the old 'Custom' tag.

Need and ReleaseOfInformation are nice additions.

What is the process and time frame to get this out of draft status?

David King
HMIS Systems Administrator SC
843-484-0257

> signature.asc
> 1KDownload

[Eric Jahn]

On Wed, 2007-08-22 at 19:51 +0000, David King wrote:
> Not sure why you left identifierStampGroup off Gender?

David, thanks for the feedback. It is very appreciated! Gender is now
fixed in the latest version in the repo. Purely an oversight, sorry.

> I also like the restructuring so that it can more easily be modfied.
> A much better replacement to the old 'Custom' tag.

Yes, 'Custom' is now fully removed.

> Need and ReleaseOfInformation are nice additions.

Good, I hope HMIS software providers agree also, but I'm open to any
ideas as to how to make Needs/ROI more useful, yet still not create
problems for developers.

> What is the process and time frame to get this out of draft status?

We don't have a set time frame, but I the march to completion is
underway. We need to get a more feedback from vendors/CoCs/implementers
before finalizing the spec. Then we need to document all the changes
from 2.7 and update all the preexisting documentation. Then it gets
submitted to HUD for approval. That process can take a while, since
they may suggest revisions (especially to the documentation), then,
voila, it's officially released on the hmis.info site. Because of some
critical extensibility improvements, it may make more sense to develop
against this draft 2.8 instead of official 2.7. Thanks again David!
Any more reviews/suggestions from folks?

-Eric