Re: [xliff] xLiff and Interoperability

Wouldn't every parser parse everything in a given XLIFF file 
anyway?  

 

The elements a given tool chooses to display to a user is 
pretty much context dependent.  So following on from Enda's comment can we 
say that "if a given attribute is in the XLIFF then we should simply 
require that a tools implementation parse it , but that we don't require the 
tool to have specific methods to operate upon it".  

 

Tools that don't process a given tag, entity or attribute 
should be obliged to write out exactly what they read in if they don't operate 
on the content specifically!

Steve.

 

 

 

 

S  t  e  p  h  e  n     
H  o  l  m  e  s
Localisation Development Manager 

International Product Development

 

Voice:  +353 (1) 241 5732
Fax:     +353 (1) 241 
5749

 

Novell, Inc., a leading provider of Net business solutions
http://www.novell.com
>>> Enda 
McDonnell <> 03/06/02 06:09 >>> 
Hi 
All, 

xLiff is young at this stage and perhaps the interoperability 
Eric asked 
about has not yet been proven; many tools have not had time to 
endorse it 
fully. However, the use of TMX between tools is a good example of 
the 
interoperability that can be achieved with xLiff as it is. 

TMX, 
which has many of the same interoperability limitations we are 
discussing 
wrt xLiff 1.0, has been used to great effect between tools. 
There are many 
examples where localisation data has been transferred between 
tools. The 
stronger commercially available localisation tools now support 
TMX. While 
that support is to differing levels, it proves that the tools 
can work with 
the flexibility that has been built into TMX and the current 
xLiff spec. 

Regarding interoperability, there are certain key pieces of data 
such as 
source and target text, parts of speech, translator notes, etc that 
are 
critical to all stages of the localisation process and need to be fully 

understandable by all consumers of xliff. There is other data that is 

completely specific to the author of the xliff file and is irrelevant to 

other users. Examples might be database unique keys from where the text 

came, or specific workflow paths within an organisation. These do not need 

to be understood by everyone, yet absolutely need to be catered for in this 

format. 

I don't believe that a flexible attribute such as 'ts', 
which xliff uses for 
this type of information, prevents interoperability. 
This may not be the 
best way of providing flexibility, but we need to 
understand that if this 
format is supposed to span the entire localisation 
process for any 
organisation, we need to provide a mechanism that allows 
users include 
proprietory information that is simply maintained by consuming 
applications 
that do not understand it. 

I am very much in favour of 
interoperability and tightly specifying as much 
as possible in xLiff 1.1, 
but I feel some flexibiltiy would promote the 
format greater adoption in the 
industry. 

Regards, 
Enda 

-----Original 
Message----- 
From: Friedman, Eric [ mailto:] 

Sent: 06 March 2002 15:16 
To: 'Peter Reynolds '; '  ' 

Subject: RE: [xliff] Minutes of the XLIFF TC March 5th. 2002 

Hi 
folks, 

On Wed, 6 Mar 2002, Peter Reynolds wrote: 

> I share 
your frustration with people describing the current spec as 
> unworkable. 
We are also working with XLIFF and doing fine on the current 
> spec. 
Considerable work went into achieving the 1.0 spec and I don't think 
> it 
is helpful for that to be dismissed. 

No one has "dismissed" the work 
that has been done up to date. I can 
only speak for myself on this subject, 
but if I thought the current 
document was worthless, I wouldn't have 
bothered to get involved. 

I've heard from several people that they have 
working implementations 
of the current specification. I think that's a 
terrific indicator that 
things are going well. However, I have >not< 
heard from anyone who 
has successfully done work using their tools in 
combination with 
those produced by someone else in the group without access 
to information 
beyond the contents of the shared spec. 

The abstract 
of the spec is very clear on this point: "The purpose of 
this format is to 
store localisable data and carry it from one step of 
the localisation 
process to the other, while allowing interoperability 
between tools." 

Has someone achieved that "interoperability between tools"? That's 

what I want to see happen, because it has much more value than 
yet 
another format that only works with tools from this-or-that 
vendor. 

Eric 

---------------------------------------------------------------- 
To 
subscribe or unsubscribe from this elist use the subscription 
manager: 
< http://lists.oasis-open.org/ob/adm.pl 
> 

---------------------------------------------------------------- 

To subscribe or unsubscribe from this elist use the subscription 

manager: < http://lists.oasis-open.org/ob/adm.pl 
>