treesoft-treefam

[Treesoft-treefam] API problem with malformed or fake trees

[Sebastien M. <seb...@un...>]

Hi,

the API should return the same thing than the web site for 'Bad Tree' 
like TF352160.

It should be clearer for users.

What do you think about it ?

-- 
Sébastien Moretti
Department of Ecology and Evolution,
Biophore, University of Lausanne,
CH-1015 Lausanne, Switzerland
Tel.: +41 (21) 692 4221/4056
http://bioinfo.unil.ch/

Re: [Treesoft-treefam] API problem with malformed or fake trees

[Jean-Karim H. <jk...@sa...>]

Hi Sebastien,

The current behaviour is to die on bad trees. I think this is the safest
procedure as not doing so would allow further processing of bad trees.
This would generate errors difficult to trace both in other parts of the
API and in any code using the API.
However, the error message could be more explicit and report on the
family AC and type of tree causing the error. Would you also have the
message say 'bad tree' instead of 'unrecognized format' ?

J-K

On Tue, 2008-10-21 at 12:15 +0200, Sebastien MORETTI wrote:
> Hi,
> 
> the API should return the same thing than the web site for 'Bad Tree' 
> like TF352160.
> 
> It should be clearer for users.
> 
> What do you think about it ?
> 

Re: [Treesoft-treefam] API problem with malformed or fake trees

[Sebastien M. <Seb...@un...>]

> Hi Sebastien,
> 
> The current behaviour is to die on bad trees. I think this is the safest
> procedure as not doing so would allow further processing of bad trees.
> This would generate errors difficult to trace both in other parts of the
> API and in any code using the API.
> However, the error message could be more explicit and report on the
> family AC and type of tree causing the error. Would you also have the
> message say 'bad tree' instead of 'unrecognized format' ?
> 
> J-K

Yes it is safer to die on bad trees.

The 'bad tree' message would be more consistent with the web site message.

Thanks
Sébastien

> On Tue, 2008-10-21 at 12:15 +0200, Sebastien MORETTI wrote:
>> Hi,
>>
>> the API should return the same thing than the web site for 'Bad Tree' 
>> like TF352160.
>>
>> It should be clearer for users.
>>
>> What do you think about it ?

-- 
Sébastien Moretti