tei-odds — Discussion of Pure ODD and its evolution

[Tei-odds] formatting pure odd specs

[Lou B. <lou...@re...>]

The Guidelines HTML pages as currently produced don't do a good job of 
handling the changes introduced with Pure ODD. In particular, here's a 
(not necessarily  complete) list of the things that ought to happen (in 
my view) but currently don't...

1. The <datatype> component of an attribute spec (whether in an 
attribute class or an element spec) currently produces some text saying  
"n to m occurrences of" (depending on the values of minOccurs and 
maxOccurs) , followed by an empty white box, optionally followed by the 
phrase "separated by whitespace". If you toggle the display of the empty 
white box from "compact" to "XML syntax", it will show the full content 
of the <content> of the <datatype> e.g.

<datatypemaxOccurs="unbounded">
<dataRefkey="teidata.pointer"/></datatype>

  The box is redundant, I think. It should be replaced simply by the 
@key value from the <dataRef> which should be an HTML link to the 
appropriate teidata spec.

2. The "Declaration" component of an elementSpec currently  (still!) 
displays its RelaxNg declaration, in either compact or XML syntax. I 
think this should be replaced by the pureODD content, displayed in the 
same way as an XML example. There is an argument for retaining the 
Relaxng declarations as well, as there is for also providing the XSD or 
DTD declarations, but I ain't making either of them.

3. Individual specs for the teidata definitions need to be generated and 
included in the lists of datatypes.

I have been poking around in the stylesheets trying to see how to make 
these things happen, and have so far (more or less) succeeded in 
achieving no 3 above. The others are more of a challenge so offers of 
help would be much appreciated.

Re: [Tei-odds] the name and attrs of <sequence>

[Lou B. <lou...@re...>]

On 26/05/15 15:54, Syd Bauman wrote:
>
> I distinctly remember being assured at the last face-to-face Council
> meeting (by Lou no less) that although Pure ODD was in the
> repository, because it was not referenced from the source files
> (like, e.g., ST-Infrastructure.xml), it would not be part of the
> released Guidelines.

My recollection of this remark was that it related only to the new 
features of Pure ODD currently under discussion (dataRef etc), not the 
other parts of pure ODD which have been included in releases since 2.60 
(see http://www.tei-c.org/Vault/P5/current/doc/tei-p5-doc/readme-2.6.0.html)

Sorry if that was not clear to slugs, attentive or otherwise.

> ...

>   And it is not at all obvious to me that the
> benefits of changing these names is sufficiently compelling to break
> backward compatibility.[1] In fact, I'm inclined to say it's not.


Good. Let's not do it then!

Re: [Tei-odds] the name and attrs of <sequence>

[Lou B. <lou...@re...>]

I agree. This is a waste of time which could be more profitably spent 
thinking about the (few) unresolved issues concerning pure ODD, i.e. how 
to do datatypes.



On 26/05/15 18:29, Sebastian Rahtz wrote:
> It's been this way for several years with no murmurings or complaint by users or class attendees. I would say birnbaum rules apply.
>
> Sebastian
> ------------------------------------------------------------------------------
> One dashboard for servers and applications across Physical-Virtual-Cloud
> Widest out-of-the-box monitoring support with 50+ applications
> Performance metrics, stats and reports that give you Actionable Insights
> Deep dive visibility with transaction tracing using APM Insight.
> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
> _______________________________________________
> tei-odds mailing list
> tei...@li...
> https://lists.sourceforge.net/lists/listinfo/tei-odds

Re: [Tei-odds] the name and attrs of <sequence>

[Sebastian R. <seb...@gm...>]

It's been this way for several years with no murmurings or complaint by users or class attendees. I would say birnbaum rules apply. 

Sebastian

Re: [Tei-odds] the name and attrs of <sequence>

[Martin H. <mh...@uv...>]

No, I'm wrong:

<http://www.tei-c.org/release/doc/tei-p5-doc/en/html/TD.html#DEFCON>

It's out there all right.

But we can presumably assume that anyone using it is doing so 
experimentally, surely, since it's not really finished; and they'll be 
subscribers to TEI-L, so we can poll them for responses to the idea of 
changing names.

Cheers,
Martin

On 15-05-26 08:08 AM, Martin Holmes wrote:
> I have been using Pure ODD a bit, but I think I'm one of a very small
> number of early adopters; I think it would be perfectly OK to take the
> course you propose to rename elements and attributes if we need to.
>
> Is there anything in the chapter prose yet? I don't see anything on a
> quick look.
>
> Cheers,
> Martin
>
> On 15-05-26 07:54 AM, Syd Bauman wrote:
>> OMG!
>>
>> I distinctly remember being assured at the last face-to-face Council
>> meeting (by Lou no less) that although Pure ODD was in the
>> repository, because it was not referenced from the source files
>> (like, e.g., ST-Infrastructure.xml), it would not be part of the
>> released Guidelines.
>>
>> But a quick look shows that it's there. I've been an inattentive slug
>> and had not noticed.
>>
>> So while I think we've made a mistake in the naming of these elements
>> (and Martin is absolutely right -- the names matter), Birnbaum
>> clearly does apply. And it is not at all obvious to me that the
>> benefits of changing these names is sufficiently compelling to break
>> backward compatibility.[1] In fact, I'm inclined to say it's not.
>>
>> Notes
>> -----
>> [1] Actually, if we were going to make this change, we would do so
>>       w/o actually breaking backward compatibility. We would do
>>       something like:
>>        1) add new elements that duplicate functionality of existing ones
>>        2) deprecate existing ones
>>        3) remove existing ones
>>       So while it doesn't actually break backward compatibility, it
>>       takes a long time and is a lot of effort. (Remember, the
>>       stylesheets have to match the Guidelines, so for awhile they
>>       would have to process both.)
>>
>> SB> Whoa! What's Birnbaum got to do with it? (This is not released
>> SB> stuff, yet.)
>>
>> LB> Au contraire...
>> LB> see eg
>> LB> http://www.tei-c.org/release/doc/tei-p5-doc/en/html/ref-alternate.html
>>
>> SR> It's been in at least the last two releases of P5, unless I am
>> SR> going mad. So I am not sure what would prompt us to change names?
>>
>> ------------------------------------------------------------------------------
>> One dashboard for servers and applications across Physical-Virtual-Cloud
>> Widest out-of-the-box monitoring support with 50+ applications
>> Performance metrics, stats and reports that give you Actionable Insights
>> Deep dive visibility with transaction tracing using APM Insight.
>> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
>> _______________________________________________
>> tei-odds mailing list
>> tei...@li...
>> https://lists.sourceforge.net/lists/listinfo/tei-odds
>>
>
> ------------------------------------------------------------------------------
> One dashboard for servers and applications across Physical-Virtual-Cloud
> Widest out-of-the-box monitoring support with 50+ applications
> Performance metrics, stats and reports that give you Actionable Insights
> Deep dive visibility with transaction tracing using APM Insight.
> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
> _______________________________________________
> tei-odds mailing list
> tei...@li...
> https://lists.sourceforge.net/lists/listinfo/tei-odds
>

Re: [Tei-odds] the name and attrs of <sequence>

[Martin H. <mh...@uv...>]

I have been using Pure ODD a bit, but I think I'm one of a very small 
number of early adopters; I think it would be perfectly OK to take the 
course you propose to rename elements and attributes if we need to.

Is there anything in the chapter prose yet? I don't see anything on a 
quick look.

Cheers,
Martin

On 15-05-26 07:54 AM, Syd Bauman wrote:
> OMG!
>
> I distinctly remember being assured at the last face-to-face Council
> meeting (by Lou no less) that although Pure ODD was in the
> repository, because it was not referenced from the source files
> (like, e.g., ST-Infrastructure.xml), it would not be part of the
> released Guidelines.
>
> But a quick look shows that it's there. I've been an inattentive slug
> and had not noticed.
>
> So while I think we've made a mistake in the naming of these elements
> (and Martin is absolutely right -- the names matter), Birnbaum
> clearly does apply. And it is not at all obvious to me that the
> benefits of changing these names is sufficiently compelling to break
> backward compatibility.[1] In fact, I'm inclined to say it's not.
>
> Notes
> -----
> [1] Actually, if we were going to make this change, we would do so
>      w/o actually breaking backward compatibility. We would do
>      something like:
>       1) add new elements that duplicate functionality of existing ones
>       2) deprecate existing ones
>       3) remove existing ones
>      So while it doesn't actually break backward compatibility, it
>      takes a long time and is a lot of effort. (Remember, the
>      stylesheets have to match the Guidelines, so for awhile they
>      would have to process both.)
>
> SB> Whoa! What's Birnbaum got to do with it? (This is not released
> SB> stuff, yet.)
>
> LB> Au contraire...
> LB> see eg
> LB> http://www.tei-c.org/release/doc/tei-p5-doc/en/html/ref-alternate.html
>
> SR> It's been in at least the last two releases of P5, unless I am
> SR> going mad. So I am not sure what would prompt us to change names?
>
> ------------------------------------------------------------------------------
> One dashboard for servers and applications across Physical-Virtual-Cloud
> Widest out-of-the-box monitoring support with 50+ applications
> Performance metrics, stats and reports that give you Actionable Insights
> Deep dive visibility with transaction tracing using APM Insight.
> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
> _______________________________________________
> tei-odds mailing list
> tei...@li...
> https://lists.sourceforge.net/lists/listinfo/tei-odds
>

Re: [Tei-odds] the name and attrs of <sequence>

[Hugh C. <phi...@gm...>]

Sigh. I remember that too…I guess it’s time for me to urge once again that we start enforcing proper branching disciple for large new features, preferably using a source control mechanism that has proper branches.

How bad of a violation of Birnbaum would this be? I’m slightly dismayed that backwards compatibility can be invoked to block changes to something still under development...

> On May 26, 2015, at 10:54 , Syd Bauman <sy...@pa...> wrote:
> 
> OMG!
> 
> I distinctly remember being assured at the last face-to-face Council
> meeting (by Lou no less) that although Pure ODD was in the
> repository, because it was not referenced from the source files
> (like, e.g., ST-Infrastructure.xml), it would not be part of the
> released Guidelines.
> 
> But a quick look shows that it's there. I've been an inattentive slug
> and had not noticed.
> 
> So while I think we've made a mistake in the naming of these elements
> (and Martin is absolutely right -- the names matter), Birnbaum
> clearly does apply. And it is not at all obvious to me that the
> benefits of changing these names is sufficiently compelling to break
> backward compatibility.[1] In fact, I'm inclined to say it's not.
> 
> Notes
> -----
> [1] Actually, if we were going to make this change, we would do so
>    w/o actually breaking backward compatibility. We would do
>    something like:
>     1) add new elements that duplicate functionality of existing ones
>     2) deprecate existing ones
>     3) remove existing ones
>    So while it doesn't actually break backward compatibility, it
>    takes a long time and is a lot of effort. (Remember, the
>    stylesheets have to match the Guidelines, so for awhile they
>    would have to process both.)
> 
> SB> Whoa! What's Birnbaum got to do with it? (This is not released
> SB> stuff, yet.)
> 
> LB> Au contraire...
> LB> see eg
> LB> http://www.tei-c.org/release/doc/tei-p5-doc/en/html/ref-alternate.html
> 
> SR> It's been in at least the last two releases of P5, unless I am
> SR> going mad. So I am not sure what would prompt us to change names?
> 
> ------------------------------------------------------------------------------
> One dashboard for servers and applications across Physical-Virtual-Cloud 
> Widest out-of-the-box monitoring support with 50+ applications
> Performance metrics, stats and reports that give you Actionable Insights
> Deep dive visibility with transaction tracing using APM Insight.
> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
> _______________________________________________
> tei-odds mailing list
> tei...@li...
> https://lists.sourceforge.net/lists/listinfo/tei-odds

Re: [Tei-odds] the name and attrs of <sequence>

[Syd B. <sy...@pa...>]

OMG!

I distinctly remember being assured at the last face-to-face Council
meeting (by Lou no less) that although Pure ODD was in the
repository, because it was not referenced from the source files
(like, e.g., ST-Infrastructure.xml), it would not be part of the
released Guidelines.

But a quick look shows that it's there. I've been an inattentive slug
and had not noticed.

So while I think we've made a mistake in the naming of these elements
(and Martin is absolutely right -- the names matter), Birnbaum
clearly does apply. And it is not at all obvious to me that the
benefits of changing these names is sufficiently compelling to break
backward compatibility.[1] In fact, I'm inclined to say it's not.

Notes
-----
[1] Actually, if we were going to make this change, we would do so
    w/o actually breaking backward compatibility. We would do
    something like:
     1) add new elements that duplicate functionality of existing ones
     2) deprecate existing ones
     3) remove existing ones
    So while it doesn't actually break backward compatibility, it
    takes a long time and is a lot of effort. (Remember, the
    stylesheets have to match the Guidelines, so for awhile they
    would have to process both.)

SB> Whoa! What's Birnbaum got to do with it? (This is not released
SB> stuff, yet.)

LB> Au contraire...
LB> see eg
LB> http://www.tei-c.org/release/doc/tei-p5-doc/en/html/ref-alternate.html

SR> It's been in at least the last two releases of P5, unless I am
SR> going mad. So I am not sure what would prompt us to change names?

Re: [Tei-odds] the name and attrs of <sequence>

[Sebastian R. <seb...@gm...>]

Sent from my iPhone

> On 23 May 2015, at 20:47, Syd Bauman <sy...@pa...> wrote:
> 
> Whoa! What's Birnbaum got to do with it? 

It's been in at least the last two releases of P5, unless I am going mad. So I am not sure what would prompt us to change names?

Sebastian 

Re: [Tei-odds] the name and attrs of <sequence>

[Lou B. <lou...@re...>]

On 23/05/15 20:47, Syd Bauman wrote:
>
> Whoa! What's Birnbaum got to do with it? (This is not released stuff,
> yet.)

Au contraire...

see eg 
http://www.tei-c.org/release/doc/tei-p5-doc/en/html/ref-alternate.html

Re: [Tei-odds] the name and attrs of <sequence>

[Syd B. <sy...@pa...>]

MH> I'd be happier with <sequence>, <bag> and <oneOf>.

Oh, I *like* that. Why not <oneOf>, <eachOf preserveOrder="true">, and
<eachOf preserveOrder="false">?

(Yes, I went back from @order to @preserveOrder -- with the GI of
<eachOf> that attribute seems clearer to me. Thoughts?)


SR> Sure. if you like. Can you justify the change at the Court of
SR> Birnbaum, though?

Whoa! What's Birnbaum got to do with it? (This is not released stuff,
yet.)

Re: [Tei-odds] the name and attrs of <sequence>

[Sebastian R. <seb...@gm...>]

> On 19 May 2015, at 21:07, Martin Holmes <mh...@uv...> wrote:
> 
> I think (as with the Simple stuff) you might be severely underestimating 
> the importance of the names. If the names of the elements are 
> transparent and comprehensible at first glance, then the whole system 
> seems easier and people will be more willing to adopt it and relinquish 
> other, harder approaches. If the names are confusing, it'll really act 
> as a barrier to adoption.

yes and no. Simple is not the same, cos it is all out on the operating table at the
moment, so the names can afford to be fluid. we’ll test next week whether the
names are bad, but we have to agree on something for testing. 

as for the sequence and alternate names, I think Lou (who has taught this
quite a few times now) would agree that the names are understood easily
by the intended audience. I don’t recall anyone being puzzled by them
in real life. For serious computer science folks, the names are
not quite right, I agree. cf Faq 11 at http://www.tvgohome.com/faq.html

Sebastian

Re: [Tei-odds] the name and attrs of <sequence>

[Sebastian R. <seb...@gm...>]

> On 19 May 2015, at 21:05, Martin Holmes <mh...@uv...> wrote:
> 
> It may be just me, but I've always found <alternate> confusing. I often 
> find myself thinking it means what I think Syd means by <bag> (selection 
> of contained items in any order), when I suppose it really means <oneOf>.
> 
> I'd be happier with <sequence>, <bag> and <oneOf>.

Sure. if you like. Can you justify the change at the Court of Birnbaum, though?

Sebastian

Re: [Tei-odds] the name and attrs of <sequence>

[Martin H. <mh...@uv...>]

On 15-05-19 01:00 PM, Sebastian Rahtz wrote:
>
> I could see the point in having both <bag> and <sequence>, and skipping the attribute.
>
> but the syntax we have was agreed quite a long time ago now, and I don’t see any great reason
> to reopen the subject of the naming, it seems, y’know, rather low priority.

I think (as with the Simple stuff) you might be severely underestimating 
the importance of the names. If the names of the elements are 
transparent and comprehensible at first glance, then the whole system 
seems easier and people will be more willing to adopt it and relinquish 
other, harder approaches. If the names are confusing, it'll really act 
as a barrier to adoption.

Cheers,
Martin

> Sebastian
> ------------------------------------------------------------------------------
> One dashboard for servers and applications across Physical-Virtual-Cloud
> Widest out-of-the-box monitoring support with 50+ applications
> Performance metrics, stats and reports that give you Actionable Insights
> Deep dive visibility with transaction tracing using APM Insight.
> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
> _______________________________________________
> tei-odds mailing list
> tei...@li...
> https://lists.sourceforge.net/lists/listinfo/tei-odds
>

Re: [Tei-odds] the name and attrs of <sequence>

[Martin H. <mh...@uv...>]

It may be just me, but I've always found <alternate> confusing. I often 
find myself thinking it means what I think Syd means by <bag> (selection 
of contained items in any order), when I suppose it really means <oneOf>.

I'd be happier with <sequence>, <bag> and <oneOf>.

Cheers,
Martin

On 15-05-19 12:33 PM, Lou Burnard wrote:
> The name sequence was chosen to distinguish this from alternate. In a
> sequence all the child elements appear in a model, in an alternate only
> one of them. We could have chosen something like catenate or string i
> suppose but sequence seems more natural. And inevitably xml being xml
> the sequence with which child elements are presented will be preserved
> so the name doesn't seem misleading to me. The attribute just tells you
> whether or not you care that the order is preserved.
>
>
>
>
> Sent from my Samsung Galaxy Tab®|PRO
>
>
> -------- Original message --------
> From: sy...@pa...
> Date:05/19/2015 20:04 (GMT+00:00)
> To: TEI ODD discussion list
> Subject: [Tei-odds] the name and attrs of
>
> Some musing on representation of content models that I began at the
> end of April, but didn't finish ... well, I suppose it's not really
> finished yet. Some of the following may be very insightful and
> helpful, or it may be complete hogwash.
>
> I finally figured out why the @preserveOrder of <sequence> bothers
> me. First, it's because we already have @order of <listChange> which
> (it seems to me) means almost exactly the same thing.
>
> Second, unless @order is set to true, it's not a sequence. Since
> items can be repeated,[1] and (unless @order is true) order does not
> matter, it's a bag by our own definitions of such things. (See 18.7,
> #FSSS.)
>
> And yes, I realize a content model is not a feature structure, but
> still, why not re-use what we already have? And for that matter,
> those of us used to XSLT will get confused in our old age when a TEI
> sequence is unordered but an XSLT sequence is ordered. So I'm
> suggesting that we use either
>
>     <bag>  for what is now <sequence preserveOrder=false>
>     <sequence>  for what is now <sequence preserveOrder=true>
>
> OR
>
>    <XXX order="true|false">
>
> where the GI here denoted "XXX" is <mayContain> or <refGrp> or <grp>
> or <formedBy> or <contains> or some such.
>
> Notes
> -----
> [1] I.e., the content model ( a, b, a ) is allowed. Of course, ( a,
>      b?, a ) is also allowed, but will cause DTDs and maybe XSDs to
>      fall over.
>
> ------------------------------------------------------------------------------
> One dashboard for servers and applications across Physical-Virtual-Cloud
> Widest out-of-the-box monitoring support with 50+ applications
> Performance metrics, stats and reports that give you Actionable Insights
> Deep dive visibility with transaction tracing using APM Insight.
> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
> _______________________________________________
> tei-odds mailing list
> tei...@li...
> https://lists.sourceforge.net/lists/listinfo/tei-odds
>
>
> ------------------------------------------------------------------------------
> One dashboard for servers and applications across Physical-Virtual-Cloud
> Widest out-of-the-box monitoring support with 50+ applications
> Performance metrics, stats and reports that give you Actionable Insights
> Deep dive visibility with transaction tracing using APM Insight.
> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
>
>
>
> _______________________________________________
> tei-odds mailing list
> tei...@li...
> https://lists.sourceforge.net/lists/listinfo/tei-odds
>

Re: [Tei-odds] the name and attrs of <sequence>

[Lou B. <lou...@re...>]

Certainly less pressing than deciding what to do about datatypes....




Sent from my Samsung Galaxy Tab®|PRO


-------- Original message --------
From: Sebastian Rahtz
Date:05/19/2015 21:00 (GMT+00:00)
To: Lou Burnard
Cc: TEI ODD discussion list
Subject: Re: [Tei-odds] the name and attrs of


I could see the point in having both <bag> and <sequence>, and skipping the attribute.

but the syntax we have was agreed quite a long time ago now, and I don’t see any great reason
to reopen the subject of the naming, it seems, y’know, rather low priority.


Sebastian

Re: [Tei-odds] the name and attrs of <sequence>

[Sebastian R. <seb...@gm...>]

I could see the point in having both <bag> and <sequence>, and skipping the attribute.

but the syntax we have was agreed quite a long time ago now, and I don’t see any great reason
to reopen the subject of the naming, it seems, y’know, rather low priority.


Sebastian

Re: [Tei-odds] the name and attrs of <sequence>

[Lou B. <lou...@re...>]

The name sequence was chosen to distinguish this from alternate. In a sequence all the child elements appear in a model, in an alternate only one of them. We could have chosen something like catenate or string i suppose but sequence seems more natural. And inevitably xml being xml the sequence with which child elements are presented will be preserved so the name doesn't seem misleading to me. The attribute just tells you whether or not you care that the order is preserved.




Sent from my Samsung Galaxy Tab®|PRO


-------- Original message --------
From: sy...@pa...
Date:05/19/2015 20:04 (GMT+00:00)
To: TEI ODD discussion list
Subject: [Tei-odds] the name and attrs of

Some musing on representation of content models that I began at the
end of April, but didn't finish ... well, I suppose it's not really
finished yet. Some of the following may be very insightful and
helpful, or it may be complete hogwash.

I finally figured out why the @preserveOrder of <sequence> bothers
me. First, it's because we already have @order of <listChange> which
(it seems to me) means almost exactly the same thing.

Second, unless @order is set to true, it's not a sequence. Since
items can be repeated,[1] and (unless @order is true) order does not
matter, it's a bag by our own definitions of such things. (See 18.7,
#FSSS.)

And yes, I realize a content model is not a feature structure, but
still, why not re-use what we already have? And for that matter,
those of us used to XSLT will get confused in our old age when a TEI
sequence is unordered but an XSLT sequence is ordered. So I'm
suggesting that we use either

   <bag>  for what is now <sequence preserveOrder=false>
   <sequence>  for what is now <sequence preserveOrder=true>

OR

  <XXX order="true|false">

where the GI here denoted "XXX" is <mayContain> or <refGrp> or <grp>
or <formedBy> or <contains> or some such.

Notes
-----
[1] I.e., the content model ( a, b, a ) is allowed. Of course, ( a,
    b?, a ) is also allowed, but will cause DTDs and maybe XSDs to
    fall over.

------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_______________________________________________
tei-odds mailing list
tei...@li...
https://lists.sourceforge.net/lists/listinfo/tei-odds

[Tei-odds] the name and attrs of <sequence>

[<sy...@pa...>]

Some musing on representation of content models that I began at the
end of April, but didn't finish ... well, I suppose it's not really
finished yet. Some of the following may be very insightful and
helpful, or it may be complete hogwash.

I finally figured out why the @preserveOrder of <sequence> bothers
me. First, it's because we already have @order of <listChange> which
(it seems to me) means almost exactly the same thing.

Second, unless @order is set to true, it's not a sequence. Since
items can be repeated,[1] and (unless @order is true) order does not
matter, it's a bag by our own definitions of such things. (See 18.7,
#FSSS.)

And yes, I realize a content model is not a feature structure, but
still, why not re-use what we already have? And for that matter,
those of us used to XSLT will get confused in our old age when a TEI
sequence is unordered but an XSLT sequence is ordered. So I'm
suggesting that we use either

   <bag>  for what is now <sequence preserveOrder=false>
   <sequence>  for what is now <sequence preserveOrder=true>

OR

  <XXX order="true|false">

where the GI here denoted "XXX" is <mayContain> or <refGrp> or <grp>
or <formedBy> or <contains> or some such.

Notes
-----
[1] I.e., the content model ( a, b, a ) is allowed. Of course, ( a,
    b?, a ) is also allowed, but will cause DTDs and maybe XSDs to
    fall over.

Re: [Tei-odds] URGENT -- new name of <textNode>

[Sebastian R. <seb...@gm...>]

since attDef/datatype can contain macroRef, and macroRef/content can
contain model.contentPart, the answer has to be yes,


Sebastian

On 22 April 2015 at 17:13, Syd Bauman <sy...@pa...> wrote:

> So can someone confirm or deny the proposition that we plan to permit
> <textNode> as a descendant of <attDef> or in the definition of an
> attribute class that might be referred to by a descendant of
> <attDef>?
>
> I.e., does Pure ODD currently allow (or will in the near future)
>
>   <attDef ident="duck">
>     <!-- the @duck attribute can have any arbitrary text -->
>     <textNode/>
>   </attDef>
>
> or
>
>   <attDef ident="quack">
>     <macroRef key="blort"/>
>   </attDef>
>   <!-- ... -->
>   <macroSpec ident="blort">
>     <content>
>       <!-- a date followed by text -->
>       <dataRef key="data.temporal.w3c"/>
>       <textNode/>
>     </content>
>   </macroSpec>
>
> ?
>
> If the answer is "no", I have no argument against <textNode>.
>
>
> > I don't remember any discussion of abolishing <textNode> during the
> > hangout, and it is not mentioned in Martin's summary. Unless
> > someone who was actually present has a different opinion, can we
> > please park this suggestion?
>
> I wasn't suggesting it *was* discussed, Lou; I was suggesting I
> *wish* it had been discussed. We don't always get what we hope for.
> :-)
>
>
> ------------------------------------------------------------------------------
> BPM Camp - Free Virtual Workshop May 6th at 10am PDT/1PM EDT
> Develop your own process in accordance with the BPMN 2 standard
> Learn Process modeling best practices with Bonita BPM through live
> exercises
> http://www.bonitasoft.com/be-part-of-it/events/bpm-camp-virtual-
> event?utm_
> source=Sourceforge_BPM_Camp_5_6_15&utm_medium=email&utm_campaign=VA_SF
> _______________________________________________
> tei-odds mailing list
> tei...@li...
> https://lists.sourceforge.net/lists/listinfo/tei-odds
>



-- 

Sebastian Rahtz

Director (Research) of Academic IT

University of Oxford IT Services

13 Banbury Road, Oxford OX2 6NN. Phone +44 1865 283431


Não sou nada.

Nunca serei nada.

Não posso querer ser nada.

À parte isso, tenho em mim todos os sonhos do mundo.

Re: [Tei-odds] URGENT -- new name of <textNode>

[Syd B. <sy...@pa...>]

So can someone confirm or deny the proposition that we plan to permit
<textNode> as a descendant of <attDef> or in the definition of an
attribute class that might be referred to by a descendant of
<attDef>?

I.e., does Pure ODD currently allow (or will in the near future) 

  <attDef ident="duck">
    <!-- the @duck attribute can have any arbitrary text -->
    <textNode/>
  </attDef>
  
or

  <attDef ident="quack">
    <macroRef key="blort"/>
  </attDef>
  <!-- ... -->
  <macroSpec ident="blort">
    <content>
      <!-- a date followed by text -->
      <dataRef key="data.temporal.w3c"/>
      <textNode/>
    </content>
  </macroSpec>

?

If the answer is "no", I have no argument against <textNode>.


> I don't remember any discussion of abolishing <textNode> during the
> hangout, and it is not mentioned in Martin's summary. Unless
> someone who was actually present has a different opinion, can we
> please park this suggestion?

I wasn't suggesting it *was* discussed, Lou; I was suggesting I
*wish* it had been discussed. We don't always get what we hope for.
:-)

Re: [Tei-odds] URGENT -- new name of <textNode>

[Lou B. <lou...@re...>]

I don't remember any discussion of abolishing <textNode> during the 
hangout, and it is not mentioned in Martin's summary. Unless someone who 
was actually present has a different opinion, can we please park this 
suggestion?


On 22/04/15 14:12, Syd Bauman wrote:
> Hold it. I thought the whole point of the Google Hangout I missed was
> to come up with a better way to say "text here" than <dataNode> or
> <textNode>, because although the latter makes perfect sense in a
> content model (where it would match what we XML folks think of as a
> text node, i.e., "text()" in XPath), it makes no sense in the
> definition of an attribute, where it is *not* a text node (at least
> not in the XPath sense).
>
> So y'all didn't actually come up with a method of referring to text?
>
>
> We could, of course, have two different mechanisms, <textNode> for
> use inside <content> models, and <dataRef key="data.text"> for use
> inside <datatype> or <attDef>. No worse than what we have now.
>
>
>> did you decide to abandon <textNode>, then? that model looks like
>>
>>   <content>
>>      <elementRef key="head" minOccurs="1" maxOccurs="1"/>
>>      <classRef key="model.glossLike"/>
>>      <textNode/>
>>    </content>
>>
>> to me. cue discussion about whether text is data, whether text == string
>> etc etc
> ------------------------------------------------------------------------------
> BPM Camp - Free Virtual Workshop May 6th at 10am PDT/1PM EDT
> Develop your own process in accordance with the BPMN 2 standard
> Learn Process modeling best practices with Bonita BPM through live exercises
> http://www.bonitasoft.com/be-part-of-it/events/bpm-camp-virtual- event?utm_
> source=Sourceforge_BPM_Camp_5_6_15&utm_medium=email&utm_campaign=VA_SF
> _______________________________________________
> tei-odds mailing list
> tei...@li...
> https://lists.sourceforge.net/lists/listinfo/tei-odds

Re: [Tei-odds] URGENT -- new name of <textNode>

[Syd B. <sy...@pa...>]

Hold it. I thought the whole point of the Google Hangout I missed was
to come up with a better way to say "text here" than <dataNode> or
<textNode>, because although the latter makes perfect sense in a
content model (where it would match what we XML folks think of as a
text node, i.e., "text()" in XPath), it makes no sense in the
definition of an attribute, where it is *not* a text node (at least
not in the XPath sense).

So y'all didn't actually come up with a method of referring to text?


We could, of course, have two different mechanisms, <textNode> for
use inside <content> models, and <dataRef key="data.text"> for use
inside <datatype> or <attDef>. No worse than what we have now.


> did you decide to abandon <textNode>, then? that model looks like
> 
>  <content>
>     <elementRef key="head" minOccurs="1" maxOccurs="1"/>
>     <classRef key="model.glossLike"/>
>     <textNode/>
>   </content>
> 
> to me. cue discussion about whether text is data, whether text == string
> etc etc

Re: [Tei-odds] URGENT -- new name of <textNode>

[Lou B. <lou...@re...>]

We didn't discuss changing textNode to anything at all. Considering the time it took to achieve consensus on that, I hope we dont. Our most recent discussion concerned a proposed data node element which we decided might be better handled by dataref.

Do please try to keep up at the back.




Sent from my Samsung Galaxy Tab®|PRO


-------- Original message --------
From: Syd Bauman
Date:04/22/2015 01:31 (GMT+00:00)
To: TEI ODD discussion list
Subject: [Tei-odds] URGENT -- new name of

What did we decide to do call <textNode>? Am I correct that it would
be <dataRef key="data.text">?

So the (bad) content model

   ( head, model.glossLike*, text )

would be expressed as

  <content>
    <elementRef key="head" minOccurs="1" maxOccurs="1"/>
    <classRef key="model.glossLike"/>
    <dataRef key="data.text"/>
  </content>

I need to know ASAP, as my handouts for a class this summer (in which
I hope to cover Pure ODD) are due, like, yesterday.

------------------------------------------------------------------------------
BPM Camp - Free Virtual Workshop May 6th at 10am PDT/1PM EDT
Develop your own process in accordance with the BPMN 2 standard
Learn Process modeling best practices with Bonita BPM through live exercises
http://www.bonitasoft.com/be-part-of-it/events/bpm-camp-virtual- event?utm_
source=Sourceforge_BPM_Camp_5_6_15&utm_medium=email&utm_campaign=VA_SF
_______________________________________________
tei-odds mailing list
tei...@li...
https://lists.sourceforge.net/lists/listinfo/tei-odds

Re: [Tei-odds] URGENT -- new name of <textNode>

[Sebastian R. <seb...@gm...>]

did you decide to abandon <textNode>, then? that model looks like

 <content>
    <elementRef key="head" minOccurs="1" maxOccurs="1"/>
    <classRef key="model.glossLike"/>
    <textNode/>
  </content>

to me. cue discussion about whether text is data, whether text == string
etc etc

On 22 April 2015 at 01:30, Syd Bauman <sy...@pa...> wrote:

> What did we decide to do call <textNode>? Am I correct that it would
> be <dataRef key="data.text">?
>
> So the (bad) content model
>
>    ( head, model.glossLike*, text )
>
> would be expressed as
>
>   <content>
>     <elementRef key="head" minOccurs="1" maxOccurs="1"/>
>     <classRef key="model.glossLike"/>
>     <dataRef key="data.text"/>
>   </content>
>
> I need to know ASAP, as my handouts for a class this summer (in which
> I hope to cover Pure ODD) are due, like, yesterday.
>
>
> ------------------------------------------------------------------------------
> BPM Camp - Free Virtual Workshop May 6th at 10am PDT/1PM EDT
> Develop your own process in accordance with the BPMN 2 standard
> Learn Process modeling best practices with Bonita BPM through live
> exercises
> http://www.bonitasoft.com/be-part-of-it/events/bpm-camp-virtual-
> event?utm_
> source=Sourceforge_BPM_Camp_5_6_15&utm_medium=email&utm_campaign=VA_SF
> _______________________________________________
> tei-odds mailing list
> tei...@li...
> https://lists.sourceforge.net/lists/listinfo/tei-odds
>



-- 

Sebastian Rahtz

Director (Research) of Academic IT

University of Oxford IT Services

13 Banbury Road, Oxford OX2 6NN. Phone +44 1865 283431


Não sou nada.

Nunca serei nada.

Não posso querer ser nada.

À parte isso, tenho em mim todos os sonhos do mundo.