key decision we made up front is that we would complement, not compete with,
the UVM information available on the Accellera site. Specifically, we decided
not to provide downloads of documents or libraries available from Accellera,
but rather to link to their download area. At least one of our competitors
provided its own downloads for Accellera content; this is permissible under the
Apache open-source licensing. However, we felt that action effectively devalued the
Accellera site, so we linked rather than
copied. This lack of overlap made it easier for Accellera first to endorse UVM
World and then to assume responsibility for its ongoing operation.
important decision we made was not to require registration to access any of the
content on the site. Since Accellera's site did not require registration for
their UVM content, it seemed appropriate for us to follow their lead. In
addition, we had observed with some other community sites that users were wary about
their profile data potentially being used for other purposes such as email
marketing. Therefore, UVM World requires registration only for posting
material, where it is necessary to prevent abuse.
recreated on UVM World two features that had proven very useful and popular on the
now-defunct OVM World site: the user forum and the community contributions
area. The UVM forum has been a
great way for novices to ask questions of the experts and for experts to
display their knowledge and offer suggestions for the evolution of the methodology.
The contributions area takes this idea one step further by
allowing experts to post documents, code, or utilities of possible interest to
the overall UVM community. Several of the enhancements made to the UVM over the
last year have been based in part on posted contributions.
already mentioned in my previous post some of the other site
features in which we made a significant investment, including the extensive
list of links to UVM content on other sites. We knew that many key players in
the UVM ecosystem had established sites where they wished to display their
expertise, so we decided that it really didn't matter whether the content or
just a link existed on UVM World. Either way, we wanted any user looking for
UVM-related information to find our site a one-stop shop. I strongly believe
that we accomplished this goal.
all else, we wanted UVM World to be a true, open community site and to be
regarded as such by the users. We disclosed on the site that it was funded by
Cadence, but we did not brand it as a Cadence property in any way. Several
other companies aligned with us in launching and endorsing the site, with more
joining in over time. Eventually, every one of our major competitors added its
logo to the ecosystem page in recognition
of UVM World's value to Accellera and to the user community. You can't get much
more open than that!
have quite a bit of background working within standards organizations,
including the IEEE, Accellera, the PCI Special Interest Group, the 1394 Trade
Association, and the now-defunct Virtual Socket Interface Alliance (VSIA). But
I can honestly say that the experience of helping to create the OVM as a de facto standard and evolve it into the
official Accellera UVM release has been my most enjoyable experience in the
world of standards. The UVM World site has had a lot to do with this successful
effort. With the donation to Accellera, I look forward to even more folks
contributing to the site to ensure that it remains the industry's UVM flagship.
truth is out there...sometimes it's in a blog.
There is a downside of the donation. The UVMe content in the UVM World website somehow disappeared. When is Accellera going to adopt UMVe and UVMsc?
Horace, the Cadence ML (multi-language) contribution is still there - http://www.uvmworld.org/contributions-details.php?id=98&keywords=UVM_ML - were you referring to something else? I can't speak for Accellera but my understanding is that ML is one of the enhancements under discussion for future UVM releases. In the meantime Cadence will continue to update and support UVM ML.