From raouljp at libero.it  Fri May  2 04:50:22 2003
From: raouljp at libero.it (=?iso-8859-1?Q?raouljp@libero.it?=)
Date: Fri,  2 May 2003 10:50:22 +0200
Subject: =?iso-8859-1?Q?acd_&_interactive_prompt?=
Message-ID: <HE94JY$35EE817150C9E2484D9968F68E1B945D@libero.it>

hi,
I need to insert interaction in my program.
How can I do it with acd, without recoding from zero the promt interaction?

tnx

by R.J.P.B.





From pmr at ebi.ac.uk  Fri May  2 06:34:56 2003
From: pmr at ebi.ac.uk (Peter Rice)
Date: Fri, 02 May 2003 11:34:56 +0100
Subject: acd & interactive prompt
References: <HE94JY$35EE817150C9E2484D9968F68E1B945D@libero.it>
Message-ID: <3EB249D0.5010705@ebi.ac.uk>

raouljp at libero.it wrote:
> hi,
> I need to insert interaction in my program.
> How can I do it with acd, without recoding from zero the promt interaction?

Interaction is not allowed in ACD.

This is a deliberate design feature to ensure that all program options 
must be set by the user (e.g. in a web form, a GUI interface, a CORBA or 
web service call) before the program starts.

Of course, you can try to interact with the user later in the program 
... and we would rather like to have, for example, interactive graphics 
displays.

There is an ajUser call that prompts the user (well, really it writes to 
standard error) and you can try to read from standard input.

Or is this just some extra interaction at startup time? Maybe ACD can 
cope with that.

Hope this helps,

Peter Rice



From raouljp at libero.it  Mon May  5 07:40:36 2003
From: raouljp at libero.it (=?iso-8859-1?Q?raouljp@libero.it?=)
Date: Mon,  5 May 2003 13:40:36 +0200
Subject: =?iso-8859-1?Q?pattern_matching_problem?=
Message-ID: <HEEWFO$027D9525023718AC0B0CE180CD49C11F@libero.it>

hi,
i need to find a pattern  in a set of sequences.
The pattern is like this one:

AATTCNATNCAACGCGAAGA

N is a base which was not be indetifyed.

regexp =  ajStrNewC("AATTCNATNCAACGCGAAGA");

match = embPatPosSeqMatchFind( seqCurr, regexp );

where seqCurr is a generic sequence of the set.
This is made for all the seqset;
the problem is, I get the match substrings from all analyzed sequences and this
 is the result :
AATTCGATGCAACGCGAAGA

"N"  seems to be a jolly char.
I need an exact match but, how  ?

by R.J.P.B.



From raouljp at libero.it  Mon May  5 10:14:11 2003
From: raouljp at libero.it (=?iso-8859-1?Q?raouljp@libero.it?=)
Date: Mon,  5 May 2003 16:14:11 +0200
Subject: =?iso-8859-1?Q?Re:pattern_matching_problem?=
Message-ID: <HEF3JN$C287FBAA85FC55724EEA9AA678AC6888@libero.it>

ok,
i have understood
:-) 

by R.J.P.B.





From jrvalverde at cnb.uam.es  Tue May  6 10:44:58 2003
From: jrvalverde at cnb.uam.es (José R. Valverde)
Date: Tue, 6 May 2003 16:44:58 +0200
Subject: pattern matching problem
In-Reply-To: <HEEWFO$027D9525023718AC0B0CE180CD49C11F@libero.it>
References: <HEEWFO$027D9525023718AC0B0CE180CD49C11F@libero.it>
Message-ID: <20030506164458.28a8b773.jrvalverde@cnb.uam.es>

On Mon,  5 May 2003 13:40:36 +0200
"raouljp at libero.it" <raouljp at libero.it> wrote:
>  is the result :
> AATTCGATGCAACGCGAAGA
> 
> "N"  seems to be a jolly char.
> I need an exact match but, how  ?
> 
Indeed! N stands for "anything" in the IUPAC codes. It is like a '*' in
most regexps.

I understand that what you want is to find sequences with actual 'N's in
them, i.e. the N is not to be substituted by its meaning, just like you 
can use 'abc\*de' in common regexps.

If you try using the program 'dreg' -or looking at its source- you'll be
able to do that: use "AATTCNATNCAACGCGAAGA" as the regexp and there you
are. Since it uses standard regexps, N is no longer interpreted as a
IUPAC code.

-- 
	These opinions are mine and only mine. Hey man, I saw them first!

			    Jos? R. Valverde

	De nada sirve la Inteligencia Artificial cuando falta la Natural


From cbeazley at hgmp.mrc.ac.uk  Fri May  9 05:20:58 2003
From: cbeazley at hgmp.mrc.ac.uk (Claude Beazley)
Date: Fri, 9 May 2003 10:20:58 +0100
Subject: coderet
Message-ID: <200305091020.58579.cbeazley@hgmp.mrc.ac.uk>

I've noticed that most emboss apps that produce a sequence file use the -outseq parameter whereas coderet seems to use -seqout shouldn't this be kept consistant?


From crowdy at bioinfo.sickkids.on.ca  Fri May  9 10:57:37 2003
From: crowdy at bioinfo.sickkids.on.ca (Edgar Crowdy)
Date: Fri, 09 May 2003 10:57:37 -0400
Subject: coderet
References: <200305091020.58579.cbeazley@hgmp.mrc.ac.uk>
Message-ID: <3EBBC1E1.3070002@bioinfo.sickkids.on.ca>

Claude Beazley wrote:
> I've noticed that most emboss apps that produce a sequence file use the -outseq parameter whereas coderet seems to use -seqout shouldn't this be kept consistant?

"seqout" is an ACD datatype, and codret is using a datatype name as a 
parameter name, which is probably not a great idea. But, as much as I'm 
all for consistency, I think you have to assume that parameter names could 
be anything.  "outseq" is a popular parameter name for output files of 
type "seqout", "seqoutall", and "seqoutset", and is actually only used 
about 20 times in EMBOSS.

-- 
Edgar Crowdy, Programmer                crowdy at bioinfo.sickkids.on.ca
Centre for Computational Biology, Hospital for Sick Children, Toronto



From pmr at ebi.ac.uk  Sun May 11 14:06:30 2003
From: pmr at ebi.ac.uk (Peter Rice)
Date: Sun, 11 May 2003 19:06:30 +0100
Subject: coderet
References: <200305091020.58579.cbeazley@hgmp.mrc.ac.uk>
Message-ID: <3EBE9126.4060506@ebi.ac.uk>

Claude Beazley wrote:
> I've noticed that most emboss apps that produce a sequence file use the -outseq parameter whereas coderet seems to use -seqout shouldn't this be kept consistant?

That's only a qualifier name ... but yes it should be consistent and 
will be one of the warning messages from acdvalid.

Peter



From =?iso-8859-1?Q?raouljp at libero.it?=  Mon May 19 08:29:39 2003
From: =?iso-8859-1?Q?raouljp at libero.it?= (=?iso-8859-1?Q?raouljp at libero.it?=)
Date: Mon, 19 May 2003 14:29:39 +0200
Subject: =?iso-8859-1?Q?Phylogenetic_tree_&_sequences?=
Message-ID: <HF4W1F$2C58B11397EE8E0F4844176562652088@libero.it>

hi,
I must associate a phylogenetic tree with a seqset.
How can i read phylogenetic tree ?
Are there library to read phylogenetic tree ?

 Using a philogenetic tree is more simple that using separate files for
different groups of sequences.
I must associate a name per group.
In this way i can select a group of sequence only by a name ( supplied by user
), and the work on them.

If there aren't library in emboss to do this, probably i'll do it by myself.

tnx
by R.J.P.B.



From jrvalverde at cnb.uam.es  Tue May 20 08:53:04 2003
From: jrvalverde at cnb.uam.es (José R. Valverde)
Date: Tue, 20 May 2003 14:53:04 +0200
Subject: Next Grant
In-Reply-To: <3EC9F3B6.4080409@ebi.ac.uk>
References: <001501c31d7d$18fc6ee0$0300000a@kuifje>
	<20030519153531.01bf48aa.jrvalverde@cnb.uam.es>
	<20030519155830.G69114-100000@fling.sanbi.ac.za>
	<20030519173248.016909d6.jrvalverde@cnb.uam.es>
	<3EC9000F.5060302@ebi.ac.uk>
	<20030520100908.253ec264.jrvalverde@cnb.uam.es>
	<3EC9F3B6.4080409@ebi.ac.uk>
Message-ID: <20030520145304.76c3b074.jrvalverde@cnb.uam.es>

We were discussing about how to facilitate/encourage people participation
in EMBOSS development and why there are so few code submissions.


On Tue, 20 May 2003 10:21:58 +0100
Peter Rice <pmr at ebi.ac.uk> wrote:
>
> So you always tell them how to contribute, right?
> 
Half as well as I'd like. And that is less than half they deserve. 
All I can tell them is to contact Alan.

Look, I'm on your side, I know, you know, _we all_ know.

What you fail to understand is that you are at _Hinxton_. Developers
outside tend to look with awe to Hinxton, where the elite dwells and
performs amazin magic. Honest.

It is very common for people outside to be shy about their work.
And everybody is told not to use a list to post a single question that
is probably a FAQ. So they don't go to emboss-dev.

Picture it another way: which looks less intimidating, an FTP server
with an "incoming" directory or one with no evident way to upload
software to? A shared disk or another guy's PC?

Or put another way: why does Jemboss ship with EMBOSS while wEMBOSS,
W2H, EMBOSS-GUI didn't? What is outsiders' perception?

Don't tell *me*. I'm on your side. I know. You know. But _they_ don't 
know.

When the TMPC addressed the core team, what we asked was for a
visible documentation of the submission process: how does one go
around it? Whom do they send their code? What is needed for code
to be accepted? If accepted, how do they gain access to the CVS
tree to maintain it?

It's no help that when we see a few guys once a year we encourage them
to go to the web site if they don't _feel_ a warm welcome there.

So, please, don't repeat what we all know. Just tell everybody what
the exact procedure is and provide an example in the website that
makes it look like "Whoa man, that was easy!! wasn't it?".

Something along the lines of "choosing a project/getting access
to a part of the CVS tree/using CVS to maintain your code/notifying
the 'core' team when it is done/getting universal recognition and a 
free beer next time we see you..." You get the idea.

And let's move this to emboss-dev from now on.

				j

--
P.S.: since you ask, I've been trying to make an Embassy for MD package
TINKER for over two years now. I work on my scarce spare time and must
reverse engineer obscure physicists-written FORTRAN code, and it takes
me time to write front-ends for over 50 programs changing over various
releases. If I ever happen to catch-up, we'll have free MD/MM code in
EMBOSS. But it's all on my own spare time, which is a real rarity, I'm
not expected to do development at work, I'm supposed to be a service
guy.

P.P.S.: on the same line, BEN developed interfaces for blast/fasta. Why,
they were shy to offer working, but localised code. I've been working on
making it generic (again on my spare time) and feel confident they should
work, sort of, but am looking for another piece of my spare time to test
my changes (amazing, isn't it?) before submitting it. You must be aware
that I need to find some time at work after I finish work and when I have
nothing else pending to simply run the test. And I'm heavily overloaded.

P.P.P.S.: AR and us are putting up a proposal to the TMPC to add LIMS
facilities to EMBOSS. *WE* care. *WE* are on your side.


From ableasby at hgmp.mrc.ac.uk  Tue May 20 11:20:58 2003
From: ableasby at hgmp.mrc.ac.uk (ableasby at hgmp.mrc.ac.uk)
Date: Tue, 20 May 2003 16:20:58 +0100 (BST)
Subject: EMBOSS code submissions
Message-ID: <200305201520.h4KFKwX20893@bromine.hgmp.mrc.ac.uk>

To answer part of Jose-Ramone's points we made the code submission
route more clear on the web pages. A new email list
(emboss-submit at embnet.org) has been added. That list reaches
the core developers who, of course, welcome submissions.

Alan


From =?iso-8859-1?Q?raouljp at libero.it?=  Wed May 21 06:09:23 2003
From: =?iso-8859-1?Q?raouljp at libero.it?= (=?iso-8859-1?Q?raouljp at libero.it?=)
Date: Wed, 21 May 2003 12:09:23 +0200
Subject: =?iso-8859-1?Q?string_reuse?=
Message-ID: <HF8EVN$58F5B76D52A82A3A6EE9001E80A74F0F@libero.it>

hi, 
i have a question about reuse of AjPStr objects.

which of the two possibilities is it correct or better?

"text" = string with a fixed text, it is used to extract substrings.

AjPStr str = NULL;
ajint i = 0;
ajint iMax = lenght_text - lenght_substring;

1) 
str = ajStrNew();
for ( i = 0; i < iMax; i++ )
 {
     ajStrAssSubC( &str, ajStrStr( text ), i, i+lenght_substring-1 );
      ... 
      ajStrDelReuse( &str);
}
ajStrDel( &str );

2) 

for ( i = 0; i < iMax; i++ )
 {
     str = ajStrNew();
     ajStrAssSubC( &str, ajStrStr( text ), i, i+lenght_substring-1 );
      ... 
      ajStrDel( &str);
}

tnx 

by R.J.P.B.
 




From pmr at ebi.ac.uk  Wed May 21 06:23:05 2003
From: pmr at ebi.ac.uk (Peter Rice)
Date: Wed, 21 May 2003 11:23:05 +0100
Subject: string reuse
References: <HF8EVN$58F5B76D52A82A3A6EE9001E80A74F0F@libero.it>
Message-ID: <3ECB5389.8060405@ebi.ac.uk>

raouljp at libero.it wrote:
> hi, 
> i have a question about reuse of AjPStr objects.
> 
> which of the two possibilities is it correct or better?
> 1) 
> str = ajStrNew();
> for ( i = 0; i < iMax; i++ )
>  {
>      ajStrAssSubC( &str, ajStrStr( text ), i, i+lenght_substring-1 );
>       ... 
>       ajStrDelReuse( &str);
> }
> ajStrDel( &str );
> 
> 2) 
> 
> for ( i = 0; i < iMax; i++ )
>  {
>      str = ajStrNew();
>      ajStrAssSubC( &str, ajStrStr( text ), i, i+lenght_substring-1 );
>       ... 
>       ajStrDel( &str);
> }

(1) Uses one string object, makes it longer if the new string is longer 
than the existing string reserved space, and deletes it at the end. it 
is very useful you expect a long string later.

You can extend this by picking a size for the string (using 
ajStrNewL(maxlength) instead of ajStrNew) to save some re-allocation.

But you do have the full length string in memory until you call ajStrDel.

(2) Creates a new string each time, and deletes it afterwards. If you 
are working with short strings it can be good.

In practice there is not much difference.

A minor philosophical point - ajStrNew does not allocate space for a 
string - it returns an extra reference-counted copy of the internal 
empty string. When you first use it, then it makes a new string.

Hope this helps,

Peter Rice



From jison at hgmp.mrc.ac.uk  Wed May 21 06:31:38 2003
From: jison at hgmp.mrc.ac.uk (Dr J.C. Ison)
Date: Wed, 21 May 2003 11:31:38 +0100
Subject: Next Grant
References: <001501c31d7d$18fc6ee0$0300000a@kuifje>
		<20030519153531.01bf48aa.jrvalverde@cnb.uam.es>
		<20030519155830.G69114-100000@fling.sanbi.ac.za>
		<20030519173248.016909d6.jrvalverde@cnb.uam.es>
		<3EC9000F.5060302@ebi.ac.uk>
		<20030520100908.253ec264.jrvalverde@cnb.uam.es>
		<3EC9F3B6.4080409@ebi.ac.uk> <20030520145304.76c3b074.jrvalverde@cnb.uam.es>
Message-ID: <3ECB558A.47B5AF00@hgmp.mrc.ac.uk>

Hi Jose

On a related note there is now a course for would-be EMBOSS developers:

"Bioinformatics Software Development using EMBOSS"
http://www.hgmp.mrc.ac.uk/CCP11/CCP11courses/EMBOSS-Course/emboss_index.html

I promoted that course to encourage EMBOSS developers outside the
EMBOSS "Inner Core" : )   All dates are booked for this year, but we've
just arranged new dates for 2004.

So if y'all know anyone who's interested, please forward the the link.

Cheers

J.


"Jos? R. Valverde" wrote:

> We were discussing about how to facilitate/encourage people participation
> in EMBOSS development and why there are so few code submissions.
>
> On Tue, 20 May 2003 10:21:58 +0100
> Peter Rice <pmr at ebi.ac.uk> wrote:
> >
> > So you always tell them how to contribute, right?
> >
> Half as well as I'd like. And that is less than half they deserve.
> All I can tell them is to contact Alan.
>
> Look, I'm on your side, I know, you know, _we all_ know.
>
> What you fail to understand is that you are at _Hinxton_. Developers
> outside tend to look with awe to Hinxton, where the elite dwells and
> performs amazin magic. Honest.
>
> It is very common for people outside to be shy about their work.
> And everybody is told not to use a list to post a single question that
> is probably a FAQ. So they don't go to emboss-dev.
>
> Picture it another way: which looks less intimidating, an FTP server
> with an "incoming" directory or one with no evident way to upload
> software to? A shared disk or another guy's PC?
>
> Or put another way: why does Jemboss ship with EMBOSS while wEMBOSS,
> W2H, EMBOSS-GUI didn't? What is outsiders' perception?
>
> Don't tell *me*. I'm on your side. I know. You know. But _they_ don't
> know.
>
> When the TMPC addressed the core team, what we asked was for a
> visible documentation of the submission process: how does one go
> around it? Whom do they send their code? What is needed for code
> to be accepted? If accepted, how do they gain access to the CVS
> tree to maintain it?
>
> It's no help that when we see a few guys once a year we encourage them
> to go to the web site if they don't _feel_ a warm welcome there.
>
> So, please, don't repeat what we all know. Just tell everybody what
> the exact procedure is and provide an example in the website that
> makes it look like "Whoa man, that was easy!! wasn't it?".
>
> Something along the lines of "choosing a project/getting access
> to a part of the CVS tree/using CVS to maintain your code/notifying
> the 'core' team when it is done/getting universal recognition and a
> free beer next time we see you..." You get the idea.
>
> And let's move this to emboss-dev from now on.
>
>                                 j
>
> --
> P.S.: since you ask, I've been trying to make an Embassy for MD package
> TINKER for over two years now. I work on my scarce spare time and must
> reverse engineer obscure physicists-written FORTRAN code, and it takes
> me time to write front-ends for over 50 programs changing over various
> releases. If I ever happen to catch-up, we'll have free MD/MM code in
> EMBOSS. But it's all on my own spare time, which is a real rarity, I'm
> not expected to do development at work, I'm supposed to be a service
> guy.
>
> P.P.S.: on the same line, BEN developed interfaces for blast/fasta. Why,
> they were shy to offer working, but localised code. I've been working on
> making it generic (again on my spare time) and feel confident they should
> work, sort of, but am looking for another piece of my spare time to test
> my changes (amazing, isn't it?) before submitting it. You must be aware
> that I need to find some time at work after I finish work and when I have
> nothing else pending to simply run the test. And I'm heavily overloaded.
>
> P.P.P.S.: AR and us are putting up a proposal to the TMPC to add LIMS
> facilities to EMBOSS. *WE* care. *WE* are on your side.

--
Jon C. Ison, PhD
Bioinformatics Applications Group
UK MRC Human Genome Mapping Project Resource Centre
Hinxton, Cambridge, CB10 1SB, UK
E-mail : jison at hgmp.mrc.ac.uk
Tel    : 01223 49-4548
HGMP-RC: http://www.hgmp.mrc.ac.uk/
EMBOSS : http://www.hgmp.mrc.ac.uk/Software/EMBOSS/





From ableasby at hgmp.mrc.ac.uk  Wed May 21 18:56:50 2003
From: ableasby at hgmp.mrc.ac.uk (ableasby at hgmp.mrc.ac.uk)
Date: Wed, 21 May 2003 23:56:50 +0100 (BST)
Subject: string reuse
Message-ID: <200305212256.h4LMuou18414@bromine.hgmp.mrc.ac.uk>


>A minor philosophical point - ajStrNew does not allocate space for a
>string - it returns an extra reference-counted copy of the internal
>empty string. When you first use it, then it makes a new string.

But that's just an implementation detail of something that is better
regarded as a more standard constructor.

Alan


From gbottu at ben.vub.ac.be  Mon May 26 05:21:42 2003
From: gbottu at ben.vub.ac.be (Guy Bottu)
Date: Mon, 26 May 2003 11:21:42 +0200 (CEST)
Subject: ACD to doc ?
Message-ID: <200305260921.h4Q9Lgsi1128222@black.vub.ac.be>

from : BEN

	Dear all,
	
I would like to write some manual pages for local developements in EMBOSS. I 
understand that there must be a tool that generates manual pages (text and 
html), or rather the sceleton of a manual page, from the ACD file. Can this tool 
be downloaded  ?

	Sincerely,
	Guy Bottu  



From raouljp at libero.it  Fri May  2 08:50:22 2003
From: raouljp at libero.it (=?iso-8859-1?Q?raouljp@libero.it?=)
Date: Fri,  2 May 2003 10:50:22 +0200
Subject: =?iso-8859-1?Q?acd_&_interactive_prompt?=
Message-ID: <HE94JY$35EE817150C9E2484D9968F68E1B945D@libero.it>

hi,
I need to insert interaction in my program.
How can I do it with acd, without recoding from zero the promt interaction?

tnx

by R.J.P.B.





From pmr at ebi.ac.uk  Fri May  2 10:34:56 2003
From: pmr at ebi.ac.uk (Peter Rice)
Date: Fri, 02 May 2003 11:34:56 +0100
Subject: acd & interactive prompt
References: <HE94JY$35EE817150C9E2484D9968F68E1B945D@libero.it>
Message-ID: <3EB249D0.5010705@ebi.ac.uk>

raouljp at libero.it wrote:
> hi,
> I need to insert interaction in my program.
> How can I do it with acd, without recoding from zero the promt interaction?

Interaction is not allowed in ACD.

This is a deliberate design feature to ensure that all program options 
must be set by the user (e.g. in a web form, a GUI interface, a CORBA or 
web service call) before the program starts.

Of course, you can try to interact with the user later in the program 
... and we would rather like to have, for example, interactive graphics 
displays.

There is an ajUser call that prompts the user (well, really it writes to 
standard error) and you can try to read from standard input.

Or is this just some extra interaction at startup time? Maybe ACD can 
cope with that.

Hope this helps,

Peter Rice



From raouljp at libero.it  Mon May  5 11:40:36 2003
From: raouljp at libero.it (=?iso-8859-1?Q?raouljp@libero.it?=)
Date: Mon,  5 May 2003 13:40:36 +0200
Subject: =?iso-8859-1?Q?pattern_matching_problem?=
Message-ID: <HEEWFO$027D9525023718AC0B0CE180CD49C11F@libero.it>

hi,
i need to find a pattern  in a set of sequences.
The pattern is like this one:

AATTCNATNCAACGCGAAGA

N is a base which was not be indetifyed.

regexp =  ajStrNewC("AATTCNATNCAACGCGAAGA");

match = embPatPosSeqMatchFind( seqCurr, regexp );

where seqCurr is a generic sequence of the set.
This is made for all the seqset;
the problem is, I get the match substrings from all analyzed sequences and this
 is the result :
AATTCGATGCAACGCGAAGA

"N"  seems to be a jolly char.
I need an exact match but, how  ?

by R.J.P.B.



From raouljp at libero.it  Mon May  5 14:14:11 2003
From: raouljp at libero.it (=?iso-8859-1?Q?raouljp@libero.it?=)
Date: Mon,  5 May 2003 16:14:11 +0200
Subject: =?iso-8859-1?Q?Re:pattern_matching_problem?=
Message-ID: <HEF3JN$C287FBAA85FC55724EEA9AA678AC6888@libero.it>

ok,
i have understood
:-) 

by R.J.P.B.





From jrvalverde at cnb.uam.es  Tue May  6 14:44:58 2003
From: jrvalverde at cnb.uam.es (José R. Valverde)
Date: Tue, 6 May 2003 16:44:58 +0200
Subject: pattern matching problem
In-Reply-To: <HEEWFO$027D9525023718AC0B0CE180CD49C11F@libero.it>
References: <HEEWFO$027D9525023718AC0B0CE180CD49C11F@libero.it>
Message-ID: <20030506164458.28a8b773.jrvalverde@cnb.uam.es>

On Mon,  5 May 2003 13:40:36 +0200
"raouljp at libero.it" <raouljp at libero.it> wrote:
>  is the result :
> AATTCGATGCAACGCGAAGA
> 
> "N"  seems to be a jolly char.
> I need an exact match but, how  ?
> 
Indeed! N stands for "anything" in the IUPAC codes. It is like a '*' in
most regexps.

I understand that what you want is to find sequences with actual 'N's in
them, i.e. the N is not to be substituted by its meaning, just like you 
can use 'abc\*de' in common regexps.

If you try using the program 'dreg' -or looking at its source- you'll be
able to do that: use "AATTCNATNCAACGCGAAGA" as the regexp and there you
are. Since it uses standard regexps, N is no longer interpreted as a
IUPAC code.

-- 
	These opinions are mine and only mine. Hey man, I saw them first!

			    Jos? R. Valverde

	De nada sirve la Inteligencia Artificial cuando falta la Natural


From cbeazley at hgmp.mrc.ac.uk  Fri May  9 09:20:58 2003
From: cbeazley at hgmp.mrc.ac.uk (Claude Beazley)
Date: Fri, 9 May 2003 10:20:58 +0100
Subject: coderet
Message-ID: <200305091020.58579.cbeazley@hgmp.mrc.ac.uk>

I've noticed that most emboss apps that produce a sequence file use the -outseq parameter whereas coderet seems to use -seqout shouldn't this be kept consistant?


From crowdy at bioinfo.sickkids.on.ca  Fri May  9 14:57:37 2003
From: crowdy at bioinfo.sickkids.on.ca (Edgar Crowdy)
Date: Fri, 09 May 2003 10:57:37 -0400
Subject: coderet
References: <200305091020.58579.cbeazley@hgmp.mrc.ac.uk>
Message-ID: <3EBBC1E1.3070002@bioinfo.sickkids.on.ca>

Claude Beazley wrote:
> I've noticed that most emboss apps that produce a sequence file use the -outseq parameter whereas coderet seems to use -seqout shouldn't this be kept consistant?

"seqout" is an ACD datatype, and codret is using a datatype name as a 
parameter name, which is probably not a great idea. But, as much as I'm 
all for consistency, I think you have to assume that parameter names could 
be anything.  "outseq" is a popular parameter name for output files of 
type "seqout", "seqoutall", and "seqoutset", and is actually only used 
about 20 times in EMBOSS.

-- 
Edgar Crowdy, Programmer                crowdy at bioinfo.sickkids.on.ca
Centre for Computational Biology, Hospital for Sick Children, Toronto



From pmr at ebi.ac.uk  Sun May 11 18:06:30 2003
From: pmr at ebi.ac.uk (Peter Rice)
Date: Sun, 11 May 2003 19:06:30 +0100
Subject: coderet
References: <200305091020.58579.cbeazley@hgmp.mrc.ac.uk>
Message-ID: <3EBE9126.4060506@ebi.ac.uk>

Claude Beazley wrote:
> I've noticed that most emboss apps that produce a sequence file use the -outseq parameter whereas coderet seems to use -seqout shouldn't this be kept consistant?

That's only a qualifier name ... but yes it should be consistent and 
will be one of the warning messages from acdvalid.

Peter



From =?iso-8859-1?Q?raouljp at libero.it?=  Mon May 19 12:29:39 2003
From: =?iso-8859-1?Q?raouljp at libero.it?= (=?iso-8859-1?Q?raouljp at libero.it?=)
Date: Mon, 19 May 2003 14:29:39 +0200
Subject: =?iso-8859-1?Q?Phylogenetic_tree_&_sequences?=
Message-ID: <HF4W1F$2C58B11397EE8E0F4844176562652088@libero.it>

hi,
I must associate a phylogenetic tree with a seqset.
How can i read phylogenetic tree ?
Are there library to read phylogenetic tree ?

 Using a philogenetic tree is more simple that using separate files for
different groups of sequences.
I must associate a name per group.
In this way i can select a group of sequence only by a name ( supplied by user
), and the work on them.

If there aren't library in emboss to do this, probably i'll do it by myself.

tnx
by R.J.P.B.



From jrvalverde at cnb.uam.es  Tue May 20 12:53:04 2003
From: jrvalverde at cnb.uam.es (José R. Valverde)
Date: Tue, 20 May 2003 14:53:04 +0200
Subject: Next Grant
In-Reply-To: <3EC9F3B6.4080409@ebi.ac.uk>
References: <001501c31d7d$18fc6ee0$0300000a@kuifje>
	<20030519153531.01bf48aa.jrvalverde@cnb.uam.es>
	<20030519155830.G69114-100000@fling.sanbi.ac.za>
	<20030519173248.016909d6.jrvalverde@cnb.uam.es>
	<3EC9000F.5060302@ebi.ac.uk>
	<20030520100908.253ec264.jrvalverde@cnb.uam.es>
	<3EC9F3B6.4080409@ebi.ac.uk>
Message-ID: <20030520145304.76c3b074.jrvalverde@cnb.uam.es>

We were discussing about how to facilitate/encourage people participation
in EMBOSS development and why there are so few code submissions.


On Tue, 20 May 2003 10:21:58 +0100
Peter Rice <pmr at ebi.ac.uk> wrote:
>
> So you always tell them how to contribute, right?
> 
Half as well as I'd like. And that is less than half they deserve. 
All I can tell them is to contact Alan.

Look, I'm on your side, I know, you know, _we all_ know.

What you fail to understand is that you are at _Hinxton_. Developers
outside tend to look with awe to Hinxton, where the elite dwells and
performs amazin magic. Honest.

It is very common for people outside to be shy about their work.
And everybody is told not to use a list to post a single question that
is probably a FAQ. So they don't go to emboss-dev.

Picture it another way: which looks less intimidating, an FTP server
with an "incoming" directory or one with no evident way to upload
software to? A shared disk or another guy's PC?

Or put another way: why does Jemboss ship with EMBOSS while wEMBOSS,
W2H, EMBOSS-GUI didn't? What is outsiders' perception?

Don't tell *me*. I'm on your side. I know. You know. But _they_ don't 
know.

When the TMPC addressed the core team, what we asked was for a
visible documentation of the submission process: how does one go
around it? Whom do they send their code? What is needed for code
to be accepted? If accepted, how do they gain access to the CVS
tree to maintain it?

It's no help that when we see a few guys once a year we encourage them
to go to the web site if they don't _feel_ a warm welcome there.

So, please, don't repeat what we all know. Just tell everybody what
the exact procedure is and provide an example in the website that
makes it look like "Whoa man, that was easy!! wasn't it?".

Something along the lines of "choosing a project/getting access
to a part of the CVS tree/using CVS to maintain your code/notifying
the 'core' team when it is done/getting universal recognition and a 
free beer next time we see you..." You get the idea.

And let's move this to emboss-dev from now on.

				j

--
P.S.: since you ask, I've been trying to make an Embassy for MD package
TINKER for over two years now. I work on my scarce spare time and must
reverse engineer obscure physicists-written FORTRAN code, and it takes
me time to write front-ends for over 50 programs changing over various
releases. If I ever happen to catch-up, we'll have free MD/MM code in
EMBOSS. But it's all on my own spare time, which is a real rarity, I'm
not expected to do development at work, I'm supposed to be a service
guy.

P.P.S.: on the same line, BEN developed interfaces for blast/fasta. Why,
they were shy to offer working, but localised code. I've been working on
making it generic (again on my spare time) and feel confident they should
work, sort of, but am looking for another piece of my spare time to test
my changes (amazing, isn't it?) before submitting it. You must be aware
that I need to find some time at work after I finish work and when I have
nothing else pending to simply run the test. And I'm heavily overloaded.

P.P.P.S.: AR and us are putting up a proposal to the TMPC to add LIMS
facilities to EMBOSS. *WE* care. *WE* are on your side.


From ableasby at hgmp.mrc.ac.uk  Tue May 20 15:20:58 2003
From: ableasby at hgmp.mrc.ac.uk (ableasby at hgmp.mrc.ac.uk)
Date: Tue, 20 May 2003 16:20:58 +0100 (BST)
Subject: EMBOSS code submissions
Message-ID: <200305201520.h4KFKwX20893@bromine.hgmp.mrc.ac.uk>

To answer part of Jose-Ramone's points we made the code submission
route more clear on the web pages. A new email list
(emboss-submit at embnet.org) has been added. That list reaches
the core developers who, of course, welcome submissions.

Alan


From =?iso-8859-1?Q?raouljp at libero.it?=  Wed May 21 10:09:23 2003
From: =?iso-8859-1?Q?raouljp at libero.it?= (=?iso-8859-1?Q?raouljp at libero.it?=)
Date: Wed, 21 May 2003 12:09:23 +0200
Subject: =?iso-8859-1?Q?string_reuse?=
Message-ID: <HF8EVN$58F5B76D52A82A3A6EE9001E80A74F0F@libero.it>

hi, 
i have a question about reuse of AjPStr objects.

which of the two possibilities is it correct or better?

"text" = string with a fixed text, it is used to extract substrings.

AjPStr str = NULL;
ajint i = 0;
ajint iMax = lenght_text - lenght_substring;

1) 
str = ajStrNew();
for ( i = 0; i < iMax; i++ )
 {
     ajStrAssSubC( &str, ajStrStr( text ), i, i+lenght_substring-1 );
      ... 
      ajStrDelReuse( &str);
}
ajStrDel( &str );

2) 

for ( i = 0; i < iMax; i++ )
 {
     str = ajStrNew();
     ajStrAssSubC( &str, ajStrStr( text ), i, i+lenght_substring-1 );
      ... 
      ajStrDel( &str);
}

tnx 

by R.J.P.B.
 




From pmr at ebi.ac.uk  Wed May 21 10:23:05 2003
From: pmr at ebi.ac.uk (Peter Rice)
Date: Wed, 21 May 2003 11:23:05 +0100
Subject: string reuse
References: <HF8EVN$58F5B76D52A82A3A6EE9001E80A74F0F@libero.it>
Message-ID: <3ECB5389.8060405@ebi.ac.uk>

raouljp at libero.it wrote:
> hi, 
> i have a question about reuse of AjPStr objects.
> 
> which of the two possibilities is it correct or better?
> 1) 
> str = ajStrNew();
> for ( i = 0; i < iMax; i++ )
>  {
>      ajStrAssSubC( &str, ajStrStr( text ), i, i+lenght_substring-1 );
>       ... 
>       ajStrDelReuse( &str);
> }
> ajStrDel( &str );
> 
> 2) 
> 
> for ( i = 0; i < iMax; i++ )
>  {
>      str = ajStrNew();
>      ajStrAssSubC( &str, ajStrStr( text ), i, i+lenght_substring-1 );
>       ... 
>       ajStrDel( &str);
> }

(1) Uses one string object, makes it longer if the new string is longer 
than the existing string reserved space, and deletes it at the end. it 
is very useful you expect a long string later.

You can extend this by picking a size for the string (using 
ajStrNewL(maxlength) instead of ajStrNew) to save some re-allocation.

But you do have the full length string in memory until you call ajStrDel.

(2) Creates a new string each time, and deletes it afterwards. If you 
are working with short strings it can be good.

In practice there is not much difference.

A minor philosophical point - ajStrNew does not allocate space for a 
string - it returns an extra reference-counted copy of the internal 
empty string. When you first use it, then it makes a new string.

Hope this helps,

Peter Rice



From jison at hgmp.mrc.ac.uk  Wed May 21 10:31:38 2003
From: jison at hgmp.mrc.ac.uk (Dr J.C. Ison)
Date: Wed, 21 May 2003 11:31:38 +0100
Subject: Next Grant
References: <001501c31d7d$18fc6ee0$0300000a@kuifje>
		<20030519153531.01bf48aa.jrvalverde@cnb.uam.es>
		<20030519155830.G69114-100000@fling.sanbi.ac.za>
		<20030519173248.016909d6.jrvalverde@cnb.uam.es>
		<3EC9000F.5060302@ebi.ac.uk>
		<20030520100908.253ec264.jrvalverde@cnb.uam.es>
		<3EC9F3B6.4080409@ebi.ac.uk> <20030520145304.76c3b074.jrvalverde@cnb.uam.es>
Message-ID: <3ECB558A.47B5AF00@hgmp.mrc.ac.uk>

Hi Jose

On a related note there is now a course for would-be EMBOSS developers:

"Bioinformatics Software Development using EMBOSS"
http://www.hgmp.mrc.ac.uk/CCP11/CCP11courses/EMBOSS-Course/emboss_index.html

I promoted that course to encourage EMBOSS developers outside the
EMBOSS "Inner Core" : )   All dates are booked for this year, but we've
just arranged new dates for 2004.

So if y'all know anyone who's interested, please forward the the link.

Cheers

J.


"Jos? R. Valverde" wrote:

> We were discussing about how to facilitate/encourage people participation
> in EMBOSS development and why there are so few code submissions.
>
> On Tue, 20 May 2003 10:21:58 +0100
> Peter Rice <pmr at ebi.ac.uk> wrote:
> >
> > So you always tell them how to contribute, right?
> >
> Half as well as I'd like. And that is less than half they deserve.
> All I can tell them is to contact Alan.
>
> Look, I'm on your side, I know, you know, _we all_ know.
>
> What you fail to understand is that you are at _Hinxton_. Developers
> outside tend to look with awe to Hinxton, where the elite dwells and
> performs amazin magic. Honest.
>
> It is very common for people outside to be shy about their work.
> And everybody is told not to use a list to post a single question that
> is probably a FAQ. So they don't go to emboss-dev.
>
> Picture it another way: which looks less intimidating, an FTP server
> with an "incoming" directory or one with no evident way to upload
> software to? A shared disk or another guy's PC?
>
> Or put another way: why does Jemboss ship with EMBOSS while wEMBOSS,
> W2H, EMBOSS-GUI didn't? What is outsiders' perception?
>
> Don't tell *me*. I'm on your side. I know. You know. But _they_ don't
> know.
>
> When the TMPC addressed the core team, what we asked was for a
> visible documentation of the submission process: how does one go
> around it? Whom do they send their code? What is needed for code
> to be accepted? If accepted, how do they gain access to the CVS
> tree to maintain it?
>
> It's no help that when we see a few guys once a year we encourage them
> to go to the web site if they don't _feel_ a warm welcome there.
>
> So, please, don't repeat what we all know. Just tell everybody what
> the exact procedure is and provide an example in the website that
> makes it look like "Whoa man, that was easy!! wasn't it?".
>
> Something along the lines of "choosing a project/getting access
> to a part of the CVS tree/using CVS to maintain your code/notifying
> the 'core' team when it is done/getting universal recognition and a
> free beer next time we see you..." You get the idea.
>
> And let's move this to emboss-dev from now on.
>
>                                 j
>
> --
> P.S.: since you ask, I've been trying to make an Embassy for MD package
> TINKER for over two years now. I work on my scarce spare time and must
> reverse engineer obscure physicists-written FORTRAN code, and it takes
> me time to write front-ends for over 50 programs changing over various
> releases. If I ever happen to catch-up, we'll have free MD/MM code in
> EMBOSS. But it's all on my own spare time, which is a real rarity, I'm
> not expected to do development at work, I'm supposed to be a service
> guy.
>
> P.P.S.: on the same line, BEN developed interfaces for blast/fasta. Why,
> they were shy to offer working, but localised code. I've been working on
> making it generic (again on my spare time) and feel confident they should
> work, sort of, but am looking for another piece of my spare time to test
> my changes (amazing, isn't it?) before submitting it. You must be aware
> that I need to find some time at work after I finish work and when I have
> nothing else pending to simply run the test. And I'm heavily overloaded.
>
> P.P.P.S.: AR and us are putting up a proposal to the TMPC to add LIMS
> facilities to EMBOSS. *WE* care. *WE* are on your side.

--
Jon C. Ison, PhD
Bioinformatics Applications Group
UK MRC Human Genome Mapping Project Resource Centre
Hinxton, Cambridge, CB10 1SB, UK
E-mail : jison at hgmp.mrc.ac.uk
Tel    : 01223 49-4548
HGMP-RC: http://www.hgmp.mrc.ac.uk/
EMBOSS : http://www.hgmp.mrc.ac.uk/Software/EMBOSS/





From ableasby at hgmp.mrc.ac.uk  Wed May 21 22:56:50 2003
From: ableasby at hgmp.mrc.ac.uk (ableasby at hgmp.mrc.ac.uk)
Date: Wed, 21 May 2003 23:56:50 +0100 (BST)
Subject: string reuse
Message-ID: <200305212256.h4LMuou18414@bromine.hgmp.mrc.ac.uk>


>A minor philosophical point - ajStrNew does not allocate space for a
>string - it returns an extra reference-counted copy of the internal
>empty string. When you first use it, then it makes a new string.

But that's just an implementation detail of something that is better
regarded as a more standard constructor.

Alan


From gbottu at ben.vub.ac.be  Mon May 26 09:21:42 2003
From: gbottu at ben.vub.ac.be (Guy Bottu)
Date: Mon, 26 May 2003 11:21:42 +0200 (CEST)
Subject: ACD to doc ?
Message-ID: <200305260921.h4Q9Lgsi1128222@black.vub.ac.be>

from : BEN

	Dear all,
	
I would like to write some manual pages for local developements in EMBOSS. I 
understand that there must be a tool that generates manual pages (text and 
html), or rather the sceleton of a manual page, from the ACD file. Can this tool 
be downloaded  ?

	Sincerely,
	Guy Bottu