Issue 3688: do we have to be more strict about the contents of  attributes                                                                                                                                          (biomolecular-ftf)
Source:  (Mr. Philip Lijnzaad, p.lijnzaad@med.uu.nl)
Nature: Uncategorized Issue
Severity: 
Summary:  do we have to be more strict about the contents of  attributes? E.g., 
    interface BioSequence
    {
        readonly attribute string             name;
        readonly attribute Identifier         id;
        readonly attribute string             description;
        readonly attribute string             seq;
        readonly attribute unsigned long      length;
        readonly attribute Basis              the_basis;

has quite a few strings, all of which could be empty. We may want to
explicitly allow  or forbid this on some of them. 

Resolution: accepted
Revised Text: Add the following sentence to 

     section 2.1.6 (Annotation, p. 2-9: return value of name) 
     section 2.1.9 (BioSequence, p. 2-23: return value of id) 
     section 2.1.18 (SearchHit, p. 2-51: return value of id). 

X shall not be empty.
Actions taken:
June 9, 2000: receieved issue
May 24, 2001: closed issue
Discussion: 

End of Annotations:=====
Date: Fri, 9 Jun 2000 15:39:29 +0100 (BST)
Message-Id: <200006091439.e59EdTA17797@o2-3.ebi.ac.uk>
X-Authentication-Warning: o2-3.ebi.ac.uk: lijnzaad set sender to
lijnzaad@ebi.ac.uk using -f
From: Philip Lijnzaad <lijnzaad@ebi.ac.uk>
To: biomolecular-ftf@emerald.omg.org
cc: muilu@ebi.ac.uk
Subject: some issues ... 
Reply-to: lijnzaad@ebi.ac.uk
Content-Type: text
X-UIDL: [C#!!<JUd9^JJ!!KKcd9

Hi all,

my apologies with being so awfully late on this, a few days before OMG
Oslo
(where I will not be going, sorry).


 do we have to be more strict about the contents of  attributes? E.g., 
    interface BioSequence
    {
        readonly attribute string             name;
        readonly attribute Identifier         id;
        readonly attribute string             description;
        readonly attribute string             seq;
        readonly attribute unsigned long      length;
        readonly attribute Basis              the_basis;

has quite a few strings, all of which could be empty. We may want to
explicitly allow  or forbid this on some of them. 

Date: Tue, 19 Sep 2000 12:34:45 +0100 (BST)
Message-Id: <200009191134.e8JBYjc43145@o2-3.ebi.ac.uk>
X-Authentication-Warning: o2-3.ebi.ac.uk: lijnzaad set sender to
lijnzaad@ebi.ac.uk using -f
From: Philip Lijnzaad <lijnzaad@ebi.ac.uk>
To: biomolecular-ftf@omg.org
cc: senger@ebi.ac.uk, muilu@ebi.ac.uk
Subject: Issue 3688: 'empty' attributes
Reply-to: lijnzaad@ebi.ac.uk
Content-Type: text
X-UIDL: 1a_!!D~Le94Kp!!N):e9


The EBI proposes to resolve issue 3688 by adding text that explicitly
forbids 
the following attributes  to be empty strings:

BioSequence::id 
BioSequence::seq to 
Annotation::name
Any occurrence of Identifier 
    (for this, add, as a first bullet to section 2.1.8.1, the sentence
    "An <bold>Identifier</bold> may not be the empty string")



                                                                      Philip
-- 
When C++ is your hammer, everything looks like a thumb. (Steven
Haflich)
-----------------------------------------------------------------------------
Philip Lijnzaad, lijnzaad@ebi.ac.uk \ European Bioinformatics
Institute,rm A2-24
+44 (0)1223 49 4639                 / Wellcome Trust Genome Campus,
Hinxton
+44 (0)1223 49 4468 (fax)           \ Cambridgeshire CB10 1SD,  GREAT
BRITAIN
PGP fingerprint: E1 03 BF 80 94 61 B6 FC  50 3D 1F 64 40 75 FB 53

Sender: muilu@ebi.ac.uk
Message-ID: <39C86E87.EDCD0FF4@ebi.ac.uk>
Date: Wed, 20 Sep 2000 08:00:08 +0000
From: Juha Muilu <muilu@ebi.ac.uk>
Reply-To: muilu@ebi.ac.uk
Organization: EMBL-EBI
X-Mailer: Mozilla 4.75 [en] (X11; U; Linux 2.2.14 i686)
X-Accept-Language: en
MIME-Version: 1.0
To: Scott Markel <smarkel@netgenics.com>
CC: lijnzaad@ebi.ac.uk, biomolecular-ftf@omg.org, senger@ebi.ac.uk
Subject: Re: Issue 3688: 'empty' attributes
References: <200009191134.e8JBYjc43145@o2-3.ebi.ac.uk>
<39C7D953.84CCEC3A@netgenics.com>
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=us-ascii
X-UIDL: Z1Ud9\1+e90;!"![^$!!

Hi,
I wonder are there cases where sequence string can be empty but
length > 0 ? 


Scott Markel wrote:
> 
> Philip,
> 
> Philip Lijnzaad wrote:
> >
> > The EBI proposes to resolve issue 3688 by adding text that
explicitly forbids
> > the following attributes  to be empty strings:
> >
> > BioSequence::id
> > BioSequence::seq to
> 
> Typo here?
> 
> > Annotation::name
> > Any occurrence of Identifier
> >     (for this, add, as a first bullet to section 2.1.8.1, the
sentence
> >     "An <bold>Identifier</bold> may not be the empty string")
> 
> I'm fine with this.  Note, however, that an NCBI-like virtual
sequence
> couldn't be represented by our BioSequence with this restriction.
> 
> Scott
> 
> --
> Scott Markel, Ph.D.       NetGenics, Inc.
> smarkel@netgenics.com     4350 Executive Drive
> Tel: 858 455 5223         Suite 260
> FAX: 858 455 1388         San Diego, CA  92121

-- 
 +--------------------------------------------------------------------+
 |Juha Muilu, Ph.D., EMBL Outstation| Email:  muilu@ebi.ac.uk
 |
 |European Bioinformatics Institute | Phone:  +44 (0)1223 494 624
 |
 |Wellcome Trust Genome Campus      | Fax:    +44 (0)1223 494 468
 |
 |Hinxton, Cambridge CB10 1SD, UK   |
 http://industry.ebi.ac.uk/~muilu|
 +--------------------------------------------------------------------+

Date: Tue, 19 Sep 2000 14:23:31 -0700
From: Scott Markel <smarkel@netgenics.com>
Organization: NetGenics, Inc.
X-Mailer: Mozilla 4.73 [en] (WinNT; U)
X-Accept-Language: en
MIME-Version: 1.0
To: lijnzaad@ebi.ac.uk
CC: biomolecular-ftf@omg.org, senger@ebi.ac.uk, muilu@ebi.ac.uk
Subject: Re: Issue 3688: 'empty' attributes
References: <200009191134.e8JBYjc43145@o2-3.ebi.ac.uk>
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=us-ascii
X-UIDL: _pY!!JWXd9+R9e9ISCe9

Philip,

Philip Lijnzaad wrote:
> 
> The EBI proposes to resolve issue 3688 by adding text that
explicitly forbids
> the following attributes  to be empty strings:
> 
> BioSequence::id
> BioSequence::seq to

Typo here?

> Annotation::name
> Any occurrence of Identifier
>     (for this, add, as a first bullet to section 2.1.8.1, the
> sentence
>     "An <bold>Identifier</bold> may not be the empty string")

I'm fine with this.  Note, however, that an NCBI-like virtual sequence
couldn't be represented by our BioSequence with this restriction.

Scott

-- 
Scott Markel, Ph.D.       NetGenics, Inc.
smarkel@netgenics.com     4350 Executive Drive
Tel: 858 455 5223         Suite 260
FAX: 858 455 1388         San Diego, CA  92121

Sender: muilu@ebi.ac.uk
Message-ID: <39C86E87.EDCD0FF4@ebi.ac.uk>
Date: Wed, 20 Sep 2000 08:00:08 +0000
From: Juha Muilu <muilu@ebi.ac.uk>
Reply-To: muilu@ebi.ac.uk
Organization: EMBL-EBI
X-Mailer: Mozilla 4.75 [en] (X11; U; Linux 2.2.14 i686)
X-Accept-Language: en
MIME-Version: 1.0
To: Scott Markel <smarkel@netgenics.com>
CC: lijnzaad@ebi.ac.uk, biomolecular-ftf@omg.org, senger@ebi.ac.uk
Subject: Re: Issue 3688: 'empty' attributes
References: <200009191134.e8JBYjc43145@o2-3.ebi.ac.uk>
<39C7D953.84CCEC3A@netgenics.com>
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=us-ascii
X-UIDL: Z1Ud9\1+e90;!"![^$!!

Hi,
I wonder are there cases where sequence string can be empty but
length > 0 ? 


Scott Markel wrote:
> 
> Philip,
> 
> Philip Lijnzaad wrote:
> >
> > The EBI proposes to resolve issue 3688 by adding text that
explicitly forbids
> > the following attributes  to be empty strings:
> >
> > BioSequence::id
> > BioSequence::seq to
> 
> Typo here?
> 
> > Annotation::name
> > Any occurrence of Identifier
> >     (for this, add, as a first bullet to section 2.1.8.1, the
sentence
> >     "An <bold>Identifier</bold> may not be the empty string")
> 
> I'm fine with this.  Note, however, that an NCBI-like virtual
sequence
> couldn't be represented by our BioSequence with this restriction.
> 
> Scott
> 
> --
> Scott Markel, Ph.D.       NetGenics, Inc.
> smarkel@netgenics.com     4350 Executive Drive
> Tel: 858 455 5223         Suite 260
> FAX: 858 455 1388         San Diego, CA  92121

-- 
 +--------------------------------------------------------------------+
 |Juha Muilu, Ph.D., EMBL Outstation| Email:  muilu@ebi.ac.uk
 |
 |European Bioinformatics Institute | Phone:  +44 (0)1223 494 624
 |
 |Wellcome Trust Genome Campus      | Fax:    +44 (0)1223 494 468
 |
 |Hinxton, Cambridge CB10 1SD, UK   |
 http://industry.ebi.ac.uk/~muilu|
 +--------------------------------------------------------------------+
Date: Wed, 20 Sep 2000 09:01:07 -0700
From: Scott Markel <smarkel@netgenics.com>
Organization: NetGenics, Inc.
X-Mailer: Mozilla 4.73 [en] (WinNT; U)
X-Accept-Language: en
MIME-Version: 1.0
To: muilu@ebi.ac.uk
CC: lijnzaad@ebi.ac.uk, biomolecular-ftf@omg.org, senger@ebi.ac.uk
Subject: Re: Issue 3688: 'empty' attributes
References: <200009191134.e8JBYjc43145@o2-3.ebi.ac.uk>
 <39C7D953.84CCEC3A@netgenics.com> <39C86E87.EDCD0FF4@ebi.ac.uk>
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=us-ascii
X-UIDL: mW%e9;mC!!~_@!!26=!!

Yes.  Quoting from NCBI documentation

  "A 'virtual' Bioseq is one in which we know the type of molecule,
  and
  possibly its length, topology, and/or strandedness, but for which we
  do not have sequence data."

Is this useful in the context of genomic maps?

Please keep in mind that I'm basically in favor of Philip's proposal.
See below.  I just want to make sure that we don't bar a particular
usage that we might want to leverage later.

Scott

Juha Muilu wrote:
> 
> Hi,
> I wonder are there cases where sequence string can be empty but
> length > 0 ?
> 
> Scott Markel wrote:
> >
> > Philip,
> >
> > Philip Lijnzaad wrote:
> > >
> > > The EBI proposes to resolve issue 3688 by adding text that
explicitly forbids
> > > the following attributes  to be empty strings:
> > >
> > > BioSequence::id
> > > BioSequence::seq to
> >
> > Typo here?
> >
> > > Annotation::name
> > > Any occurrence of Identifier
> > >     (for this, add, as a first bullet to section 2.1.8.1, the
sentence
> > >     "An <bold>Identifier</bold> may not be the empty string")
> >
> > I'm fine with this.  Note, however, that an NCBI-like virtual
sequence
> > couldn't be represented by our BioSequence with this restriction.
> >
> > Scott

-- 
Scott Markel, Ph.D.       NetGenics, Inc.
smarkel@netgenics.com     4350 Executive Drive
Tel: 858 455 5223         Suite 260
FAX: 858 455 1388         San Diego, CA  92121

From: "Dickson, Mike" <mdickson@netgenics.com>
To: "Markel, Scott" <smarkel@netgenics.com>, muilu@ebi.ac.uk
Cc: lijnzaad@ebi.ac.uk, biomolecular-ftf@omg.org, senger@ebi.ac.uk
Subject: RE: Issue 3688: 'empty' attributes
Date: Wed, 20 Sep 2000 12:28:54 -0400
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2650.21)
Content-Type: text/plain;
	      charset="iso-8859-1"
X-UIDL: n8R!!#O1e93_~e9W9od9

Why isn't this enough of a use case on its own?  I.e. isnt the fact
that you
can describe a use case for one enough?  Even in this case you do have
a
stable ID right?  Its just the sequence part that can be empty.

Mike

> -----Original Message-----
> From: Scott Markel [mailto:smarkel@netgenics.com]
> Sent: Wednesday, September 20, 2000 12:01 PM
> To: muilu@ebi.ac.uk
> Cc: lijnzaad@ebi.ac.uk; biomolecular-ftf@omg.org; senger@ebi.ac.uk
> Subject: Re: Issue 3688: 'empty' attributes
> 
> 
> Yes.  Quoting from NCBI documentation
> 
>   "A 'virtual' Bioseq is one in which we know the type of 
> molecule, and
>   possibly its length, topology, and/or strandedness, but for which
we
>   do not have sequence data."
> 
> Is this useful in the context of genomic maps?
> 
> Please keep in mind that I'm basically in favor of Philip's
proposal.
> See below.  I just want to make sure that we don't bar a particular
> usage that we might want to leverage later.
> 
> Scott
> 
> Juha Muilu wrote:
> > 
> > Hi,
> > I wonder are there cases where sequence string can be empty but
> > length > 0 ?
> > 
> > Scott Markel wrote:
> > >
> > > Philip,
> > >
> > > Philip Lijnzaad wrote:
> > > >
> > > > The EBI proposes to resolve issue 3688 by adding text 
> that explicitly forbids
> > > > the following attributes  to be empty strings:
> > > >
> > > > BioSequence::id
> > > > BioSequence::seq to
> > >
> > > Typo here?
> > >
> > > > Annotation::name
> > > > Any occurrence of Identifier
> > > >     (for this, add, as a first bullet to section 
> 2.1.8.1, the sentence
> > > >     "An <bold>Identifier</bold> may not be the empty string")
> > >
> > > I'm fine with this.  Note, however, that an NCBI-like 
> virtual sequence
> > > couldn't be represented by our BioSequence with this
restriction.
> > >
> > > Scott
> 
> -- 
> Scott Markel, Ph.D.       NetGenics, Inc.
> smarkel@netgenics.com     4350 Executive Drive
> Tel: 858 455 5223         Suite 260
> FAX: 858 455 1388         San Diego, CA  92121
> 
Date: Mon, 25 Sep 2000 13:39:12 +0100 (BST)
Message-Id: <200009251239.e8PCdCQ42726@o2-3.ebi.ac.uk>
X-Authentication-Warning: o2-3.ebi.ac.uk: lijnzaad set sender to
lijnzaad@ebi.ac.uk using -f
From: Philip Lijnzaad <lijnzaad@ebi.ac.uk>
To: smarkel@netgenics.com
CC: biomolecular-ftf@omg.org, senger@ebi.ac.uk, muilu@ebi.ac.uk
In-reply-to: <39C7D953.84CCEC3A@netgenics.com> (message from Scott
Markel on
       Tue, 19 Sep 2000 14:23:31 -0700)
Subject: Re: Issue 3688: 'empty' attributes
Reply-to: lijnzaad@ebi.ac.uk
References: <200009191134.e8JBYjc43145@o2-3.ebi.ac.uk>
<39C7D953.84CCEC3A@netgenics.com>
Content-Type: text
X-UIDL: ,h7!!20Oe9V&a!!a='!!


Scott> Philip,
Scott> Philip Lijnzaad wrote:
>> 
>> The EBI proposes to resolve issue 3688 by adding text that
explicitly forbids
>> the following attributes  to be empty strings:
>> 
>> BioSequence::id
>> BioSequence::seq to

Scott> Typo here?

the "to" is a typo, but I think we should state that BioSequence::seq
should
not be empty, simply because it would be a bit useless.

>> Annotation::name
>> Any occurrence of Identifier
>> (for this, add, as a first bullet to section 2.1.8.1, the sentence
>> "An <bold>Identifier</bold> may not be the empty string")

Scott> I'm fine with this.  Note, however, that an NCBI-like virtual
sequence
Scott> couldn't be represented by our BioSequence with this
restriction.

? Don't know them, but would argue that you always want the
aminoacids/nucleotides, whatsoever. And since this is CORBA, this can
be done
cheaply; you don't have to explicitly store all those nucleotides if
they
actually are constructed on the fly from lots of different other
sequences. 

BTW, do we state somewhere that the length attribute must be identical
to the
length of the  seq string ? I think we should. Cheers,

                                                                      Philip

-- 
When C++ is your hammer, everything looks like a thumb. (Steven
Haflich)
-----------------------------------------------------------------------------
Philip Lijnzaad, lijnzaad@ebi.ac.uk \ European Bioinformatics
Institute,rm A2-24
+44 (0)1223 49 4639                 / Wellcome Trust Genome Campus,
Hinxton
+44 (0)1223 49 4468 (fax)           \ Cambridgeshire CB10 1SD,  GREAT
BRITAIN
PGP fingerprint: E1 03 BF 80 94 61 B6 FC  50 3D 1F 64 40 75 FB 53

Date: Mon, 25 Sep 2000 14:09:31 -0700
From: Scott Markel <smarkel@netgenics.com>
Organization: NetGenics, Inc.
X-Mailer: Mozilla 4.73 [en] (WinNT; U)
X-Accept-Language: en
MIME-Version: 1.0
To: lijnzaad@ebi.ac.uk
CC: biomolecular-ftf@omg.org, senger@ebi.ac.uk, muilu@ebi.ac.uk
Subject: Re: Issue 3688: 'empty' attributes
References: <200009191134.e8JBYjc43145@o2-3.ebi.ac.uk>
<39C7D953.84CCEC3A@netgenics.com>
<200009251239.e8PCdCQ42726@o2-3.ebi.ac.uk>
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=us-ascii
X-UIDL: 9SI!!kRC!!m8j!!%>b!!

Philip,

Philip Lijnzaad wrote:
> 
> Scott> Philip Lijnzaad wrote:
> >>
> >> The EBI proposes to resolve issue 3688 by adding text that
explicitly forbids
> >> the following attributes  to be empty strings:
> >>
> >> BioSequence::id
> >> BioSequence::seq to
> 
> Scott> Typo here?
> 
> the "to" is a typo, but I think we should state that
BioSequence::seq should
> not be empty, simply because it would be a bit useless.

I disagree on the uselessness.  Quoting from NCBI documentation

  "A 'virtual' Bioseq is one in which we know the type of molecule,
  and
  possibly its length, topology, and/or strandedness, but for which we
  do not have sequence data."

I don't think we should preclude this usage.

> >> Annotation::name
> >> Any occurrence of Identifier
> >> (for this, add, as a first bullet to section 2.1.8.1, the
> >> sentence
> >> "An <bold>Identifier</bold> may not be the empty string")
> 
> Scott> I'm fine with this.  Note, however, that an NCBI-like virtual
> >> sequence
> Scott> couldn't be represented by our BioSequence with this
> >> restriction.
> 
> ? Don't know them, but would argue that you always want the
> aminoacids/nucleotides, whatsoever. And since this is CORBA, this
> >> can be done
> cheaply; you don't have to explicitly store all those nucleotides if
> >> they
> actually are constructed on the fly from lots of different other
> >> sequences.

It not a question of expense.  There may be cases where the bases just
aren't known, but there is other biosequence information that is
known.

> BTW, do we state somewhere that the length attribute must be
  identical to the
> length of the  seq string ? I think we should. Cheers,

We don't and I don't think we should.  The latter obvious follows from
my comments above.

Scott

-- 
Scott Markel, Ph.D.       NetGenics, Inc.
smarkel@netgenics.com     4350 Executive Drive
Tel: 858 455 5223         Suite 260
FAX: 858 455 1388         San Diego, CA  92121

Date: Tue, 26 Sep 2000 09:55:45 +0100 (BST)
Message-Id: <200009260855.e8Q8tjv46648@o2-3.ebi.ac.uk>
X-Authentication-Warning: o2-3.ebi.ac.uk: lijnzaad set sender to
lijnzaad@ebi.ac.uk using -f
From: Philip Lijnzaad <lijnzaad@ebi.ac.uk>
To: smarkel@netgenics.com
CC: biomolecular-ftf@omg.org, senger@ebi.ac.uk, muilu@ebi.ac.uk
In-reply-to: <39CFBF0B.9F06EAED@netgenics.com> (message from Scott
Markel on
       Mon, 25 Sep 2000 14:09:31 -0700)
Subject: Re: Issue 3688: 'empty' attributes
Reply-to: lijnzaad@ebi.ac.uk
References: <200009191134.e8JBYjc43145@o2-3.ebi.ac.uk>
<39C7D953.84CCEC3A@netgenics.com>
<200009251239.e8PCdCQ42726@o2-3.ebi.ac.uk>
<39CFBF0B.9F06EAED@netgenics.com>
Content-Type: text
X-UIDL: i+e!!<6Z!!0~Ne9~/Me9


>> >> The EBI proposes to resolve issue 3688 by adding text that
      explicitly forbids
>> >> the following attributes  to be empty strings:
>> >>
>> >> BioSequence::id
>> >> BioSequence::seq to
>> 
>> 
>> the "to" is a typo, but I think we should state that
      BioSequence::seq should
>> not be empty, simply because it would be a bit useless.

Scott> I disagree on the uselessness.  Quoting from NCBI documentation

Scott>   "A 'virtual' Bioseq is one in which we know the type of
molecule, and
Scott>   possibly its length, topology, and/or strandedness, but for
which we
Scott>   do not have sequence data."

OK, thanks, I thought they were less ephemeral. It seems I'm the only
one
wanting to have these stronger semantics, so I won't insist (but see
below). So that would leave Annotation::name and all Identifiers as
not being
allowed to be empty.

>> BTW, do we state somewhere that the length attribute must be
   identical to the
>> length of the  seq string ? I think we should. Cheers,

Scott> We don't and I don't think we should.  The latter obvious
follows from
Scott> my comments above.

So ... is this BioSequence::length <-> BioSequence::seq.length()
mismatch
only allowed if BioSequence::seq.equals("") ? If not, why not (use
case), and
what are the semantics in this case? Just to give my position: apart
from the
virtual sequence bit, I think the
two should be equal 

  - to satisfy common expectation
  - to be able to rely on the length attribute when having to manage
  - memory
    on the client side
  - to be able to decided whether to get the nucleotides 'all in one
  - go' or
    in chunks using seq_interval().

Cheers,
                                                                      Philip

-- 
When C++ is your hammer, everything looks like a thumb. (Steven
Haflich)
-----------------------------------------------------------------------------
Philip Lijnzaad, lijnzaad@ebi.ac.uk \ European Bioinformatics
Institute,rm A2-24
+44 (0)1223 49 4639                 / Wellcome Trust Genome Campus,
Hinxton
+44 (0)1223 49 4468 (fax)           \ Cambridgeshire CB10 1SD,  GREAT
BRITAIN
PGP fingerprint: E1 03 BF 80 94 61 B6 FC  50 3D 1F 64 40 75 FB 53

Date: Tue, 26 Sep 2000 10:15:37 -0700
From: Scott Markel <smarkel@netgenics.com>
Organization: NetGenics, Inc.
X-Mailer: Mozilla 4.73 [en] (WinNT; U)
X-Accept-Language: en
MIME-Version: 1.0
To: lijnzaad@ebi.ac.uk
CC: biomolecular-ftf@omg.org, senger@ebi.ac.uk, muilu@ebi.ac.uk
Subject: Re: Issue 3688: 'empty' attributes
References: <200009191134.e8JBYjc43145@o2-3.ebi.ac.uk>
<39C7D953.84CCEC3A@netgenics.com>
<200009251239.e8PCdCQ42726@o2-3.ebi.ac.uk>
<39CFBF0B.9F06EAED@netgenics.com>
<200009260855.e8Q8tjv46648@o2-3.ebi.ac.uk>
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=us-ascii
X-UIDL: mCad9<:L!!]Mn!!^S_d9

Philip,

Philip Lijnzaad wrote:
> 
> >> >> The EBI proposes to resolve issue 3688 by adding text that
explicitly forbids
> >> >> the following attributes  to be empty strings:
> >> >>
> >> >> BioSequence::id
> >> >> BioSequence::seq to
> >>
> >> the "to" is a typo, but I think we should state that
BioSequence::seq should
> >> not be empty, simply because it would be a bit useless.
> 
> Scott> I disagree on the uselessness.  Quoting from NCBI
documentation
> 
> Scott>   "A 'virtual' Bioseq is one in which we know the type of
molecule, and
> Scott>   possibly its length, topology, and/or strandedness, but for
which we
> Scott>   do not have sequence data."
> 
> OK, thanks, I thought they were less ephemeral. It seems I'm the
only one
> wanting to have these stronger semantics, so I won't insist (but see
> below). So that would leave Annotation::name and all Identifiers as
not being
> allowed to be empty.

Okay with me.

> >> BTW, do we state somewhere that the length attribute must be
     identical to the
> >> length of the  seq string ? I think we should. Cheers,
> 
> Scott> We don't and I don't think we should.  The latter obvious
     follows from
> Scott> my comments above.
> 
> So ... is this BioSequence::length <-> BioSequence::seq.length()
     mismatch
> only allowed if BioSequence::seq.equals("") ?

That's probably a reasonable restriction.

> If not, why not (use case), and
> what are the semantics in this case?

The only other realworld case is that you know/think/conjecture that a
sequence is 10K bases long, but you only know 300 bases at the 3' end
and 250 bases at the 5' end.  While this is a real example, the
current
BSA model wasn't designed to support this.  I think we'd all agree
that
what we had in mind was a single string with known bases at the core
of
a BioSequence.  An empty string is a special case of this.  The
sequence
with known ends and unknown middle should probably be made from two
sequences and, hence, is a different beast.

> Just to give my position: apart from the
> virtual sequence bit, I think the
> two should be equal
> 
>   - to satisfy common expectation
>   - to be able to rely on the length attribute when having to manage
> memory
>     on the client side
>   - to be able to decided whether to get the nucleotides 'all in one
> go' or
>     in chunks using seq_interval().

Agreed.

Scott

-- 
Scott Markel, Ph.D.       NetGenics, Inc.
smarkel@netgenics.com     4350 Executive Drive
Tel: 858 455 5223         Suite 260
FAX: 858 455 1388         San Diego, CA  92121