Yes, I couldn't say it better myself. I'm sure that people have very good
components and solutions which are now swapped by email and such. Yes Tassel
is a good place, and we have a few other places to go to, but wouldn't it
be nice if the main Tapestry website had some sanctioned space, not only in
the wiki.
I was also thinking more along the lines of maybe having a new "contrib"
section which continually grew, and then have a process where people could
contribute stuff, someone who got the ticket in a working gruop pestered the
autho untill all docs were in place, and then it was published and
downloadable from the main site (or wherever, with a link).
Then it struck me that this is the way it is today, and maybe the only thing
needed is to get more people involved in the process.
Cheers,
PS
On 5/17/06, Brian K. Wallace <[EMAIL PROTECTED]> wrote:
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
There are reasons why I'll never see that methodology in some places,
but let's stick to where it can work: Despite issues anyone may have
with Maven, why not use ibiblio for it? Instead of YAPTG (yet another
place to go)?
I'm still in favor of getting more of the commonly requested services
packaged together. Some I believe would be nice in Tapestry (not core,
but as contrib is today), some less so. I just think we need to start
getting the "componentization" (what Tapestry preaches) back into
services as well as into presentation.
Konstantin Ignatyev wrote:
>
<!-- @page { size: 8.5in 11in; margin: 0.79in } P
{ margin-bottom: 0.08in } --> Those are valid concerns Brian.
There are my thoughts on that:
> - Network is unavailable: dependency manager should maintain a local
cache of already downloaded artifacts to avoid unnecessary network
trips. Maven dependencies resolver already has that but they have bad
update policy IMO that requires rechecking with repository at least once a
day. Unnecessary IMO and bad by design because this way they try to
implement auto updates based on version ranges in the dependencies: very bad
idea because it leads to the build unpredictability by design!
> - Server is unavailable: there should be network of servers. It is very
common approach and works remarkably well (Gentoo, Ubuntoo, and other
GNU-Linux distros);
> - Speed of operations and network outages: for serious use company or
person should be able easily install a local copy of the repository within
own network that will allow (semi-)independent operation and provide place
for artifacts we or they cannot place on the public servers. Examples of
those:
> - SUNs j2ee interface jars cannot be placed in the repository
because of licensing limitations;
> - Company's private assets have no place in the public repositories
too;
>
> All the above makes me believe that 'one download' concept should an can
be retired. The repository seed can be made downloadable as DVD image - that
would make sense IMO.
>
> Konstantin Ignatyev
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.5 (MingW32)
iD8DBQFEa0xDaCoPKRow/gARAqr+AJ4ttVE9hLKCT0QOMPjbmwRcUBT8XwCfaKIa
N6LlPaDUiPbjsVFx9my9xgE=
=Qrex
-----END PGP SIGNATURE-----
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
