Profiling in v5.4

[margaret.devere]

Hi folks --

Something very problematic has just happened to my entire file set (only
about 50 files affected, but still ....)

We have a specialization that allows us to apply profiling across 9
dimensions. For certain dimensions, we apply more than one value (e.g.,
product="ProductA ProductB"). AE shows this in red strikeout, but v5.3
always went ahead and rendered it, and I had everything set up and
working.

I upgraded to v5.4 and got that working. Copied the custom folder over,
etc. etc.

Now, I'm able to apply profiles just the way I always was, but when I
render, all of the elements that had this "multiple" profile value get
their profile attribute stripped. I check the content before I render,
and all the values are there. I check after I render, and the profiles
are gone.

So I ran the Validate DITA Specialization utility and got 36 errors, all
along the lines of "architecture is an improper additional attribute"
("architecture" is one of our custom dimensions).

I read the install.pdf for v5.4 and it says "... thoroughly test the
functionality. This is especially important in the areas of ...
profiling, ...." And that's all they say.

I don't think I can use profile groups -- they would not be mutually
exclusive within a doc.

It looks to me as if we need to update our specialization. We had this
done for us so I don't know much about it but I can plunge in if need
be.

Any ideas from any angle? Something I can turn off in the rendering
process to avoid updating the specialization? I can always publish from
v5.3 but that won't work in the long run -- I'm dying to use v5.4
Styler.


Thanks,
Margaret Devere

P.S. Hello to all the folks I met at PTCUser, and thanks for the many
tips.

[naglists]

Hi Margaret,

Probably you know all this but just in case.

Arbortext profiling expects multiple values to be separated by semicolons so
it "should" be:
product="ProductA;ProductB"

And, of course, Arbortext expects the attribute to be defined in the DTD in
the element's ATTLIST:
product CDATA #IMPLIED
>

Can you change one element's definition in the DTD and one element
instance's profiling to match Arbortext's expectations and see what happens?

On Tue, Jul 14, 2009 at 2:59 PM, Margaret Devere <
margaret.devere@emerson.com> wrote:

> Hi folks --
>
> Something very problematic has just happened to my entire file set (only
> about 50 files affected, but still ....)
>
> We have a specialization that allows us to apply profiling across 9
> dimensions. For certain dimensions, we apply more than one value (e.g.,
> product="ProductA ProductB"). AE shows this in red strikeout, but v5.3
> always went ahead and rendered it, and I had everything set up and
> working.
>
> I upgraded to v5.4 and got that working. Copied the custom folder over,
> etc. etc.
>
> Now, I'm able to apply profiles just the way I always was, but when I
> render, all of the elements that had this "multiple" profile value get
> their profile attribute stripped. I check the content before I render,
> and all the values are there. I check after I render, and the profiles
> are gone.
>
> So I ran the Validate DITA Specialization utility and got 36 errors, all
> along the lines of "architecture is an improper additional attribute"
> ("architecture" is one of our custom dimensions).
>
> I read the install.pdf for v5.4 and it says "... thoroughly test the
> functionality. This is especially important in the areas of ...
> profiling, ...." And that's all they say.
>
> I don't think I can use profile groups -- they would not be mutually
> exclusive within a doc.
>
> It looks to me as if we need to update our specialization. We had this
> done for us so I don't know much about it but I can plunge in if need
> be.
>
> Any ideas from any angle? Something I can turn off in the rendering
> process to avoid updating the specialization? I can always publish from
> v5.3 but that won't work in the long run -- I'm dying to use v5.4
> Styler.
>
>
> Thanks,
> Margaret Devere
>
> P.S. Hello to all the folks I met at PTCUser, and thanks for the many
> tips.
>

[margaret.devere]

No, I didn't know any of that.



I started by adding a semicolon (wrong thing to do first), and AE still
showed this as an error.



Then I tried to find the relevant part of the relevant DTD, and it
appears that the DTD is referencing an external file that was part of
our specialization. Below is what I think is the relevant part of the
concept.dtd, and following that is the MMIFlowMeterPropsDomain content.
I don't know if this is a good way to put the profiling together.



It seems to have miraculously started working again, but I'll have to go
through and reapply all the profiling, and I don't trust it now.





From concept.dtd













MMIFlowMeterProps Domain//EN"

"MMIFlowMeterPropsDomain.ent" >

%MMIFlowMeterProps-d-dec;



.

.

.










" %architectureProps-d-attribute; %toolProps-d-attribute;
%applicationProps-d-attribute;

%approvalsProps-d-attribute; %cableProps-d-attribute;
%CPProps-d-attribute; "

>





From MMIFlowMeterProps.ent:




'props CDATA #IMPLIED

%platformProps-d-attribute;

%productProps-d-attribute;

audience CDATA #IMPLIED

otherprops CDATA #IMPLIED

%props-attribute-extensions; ' >




















"application (API | Batching | CustodyNTEP | CustodyOIML |
Concentration | FillingDosing |

MeterVerification | NOC_MVD) #IMPLIED"

>




"approvals (ATEX_Zone1 | ATEX_Zone2 ) #IMPLIED"

>




"architecture (Integral | 4wire | 9wire | DoubleHop) #IMPLIED"

>




"cable (Armored | Jacketed | Shielded) #IMPLIED"

>




"CP (Enhanced | Standard | Both) #IMPLIED"

>




"tool (AMS | DeviceNet_EDS | DeviceNet_OIA | DeviceNet_InputAsmbly |
DeviceNet_OutputAsmbly

| DeviceNet_Switches | Profibus-DP_BusParam | Profibus-DP_EDD |
Profibus-DP_GSD

| Profibus-DP_Switches | Fieldbus | 375 | Modbus | Profibus-PA_EDD |
Profibus-PA_GSD

| Profibus-PA_BusParam | ProLink-II | ProLink-III

| PocketProLink | Display | ZeroButton ) #IMPLIED"

>




"product (ELITE | F-Series | T-Series | H-Series | R-Series | CNG |
1500_Analog | 1500_FT

| 1700_Analog | 1700_IS | 2200S_Analog | 2400S_Analog |
2400S_Profibus-DP

| 2400S_DeviceNet | 2500_CIO | 2700_Analog | 2700_IS | 2700_CIO |
2700_Fieldbus

| 2700_Profibus-PA | 3300_Panel | 3300_Rack | 3350_Field | 3500_Panel


| 3500_Rack | 3700_Field | RFT9739_MVD) #IMPLIED"

>




"platform (CurvedTube | StraightTube | Controller | Transmitter)
#IMPLIED"

>










"a(props application )"

>




"a(props approvals )"

>




"a(props architecture )"

>




"a(props cable )"

>




"a(props CP )"

>


"a(props tool )"

>

[ptc-953343]

sounds like your catalog wasn't read - just a guess

[naglists]

Hi Margaret,
I can't get to a .DCF file at the moment, but profiling also needs to be
configured there (or in an external file referenced by the .DCF). If you are
specifying (or if your customization is specifying) a token list, I suspect
it must match exactly the values set up in the .DCF profiling configuration.

Our application doesn't rely on the DTD to validate those attribute entries
... we assume that the authors are going to use the profiling dialog (which
is configured / driven by the .DCF) to add values to those attributes.