Re: [ecf-dev] file-based discovery

[Scott Lewis <slewis@xxxxxxxxxxxxx>]

Hi Sebastian,

On 2/4/2011 2:10 AM, Sebastian Schmidt wrote:
> Hi,
>
> perhaps we might deprecate the old discovery mechanism in ECF 3.5 and
> remove it in ECF 3.6. This would give consumers more time to migrate
> their code.

Normally, I would say this is a fine/good idea, but there are several 
things that make this problematic:

1) Given the move to using RSA (underneath the ECF remote service 
support), it may be technically difficult (or impossible) to include 
both the old and the new/standards-based file-based discovery 
mechanisms.  That is...at this moment, I'm not even sure how much work 
it would be to make the old xml-file-based discovery work with RSA (it 
might be a lot...for example the existing/old xml-file-based code 
doesn't even compile at the moment...because it's dependent upon some of 
the old RFC 119-based classes...which have been removed in favor of the 
RSA impl).

2) It may be very confusing for people to have *both* xml-file formats 
supported in one implementation (assuming 1 can be addressed).  Because 
essentially they both do exactly the same thing (allow remote services 
to be discovered via an xml-file-based description of the remote 
service).  One (the RSA endpoint description extender format) is an 
actual standard, while the old/existing format is based upon what was 
defined for RFC 119...and that RFC was discarded at the last minute).

3) It will a fair amount of effort to support the old format, and there 
are no resources that I know of for support (for example, I simply won't 
be able to do it, given my expected availability...and the last thing 
that I want is to be dinged at the project level over and over for not 
providing sufficient documentation and support).

So, I would say that if there is someone willing to provide resources to 
deal with 1, 2, and 3 (e.g. to change the existing code to work with 
RSA, to document and support the old format...in combination with the 
new format), then we could do this strategy.

But lacking contributions to make this happen, I don't see how it's 
going to be possible...for reasons detailed above.

So...we'll put it to folks:  If you want the old format supported for 
ECF 3.5, then please contribute the resources to address the three items 
above.  Otherwise, I don't see any other way than to support the new 
RSA/OSGi standard format exclusively.

Scott