How do we improve Voldemort's community, visibility, etc.?

[Kirk True]

Hi all,

I'd like to start a non-technical discussion about Voldemort's community, visibility, and so on.

First off, I love Voldemort. I'm a committer and I respect Jay, Alex, Roshan, Bhupesh immensely. I've been extremely blessed to work with those guys on Voldemort. I'm not into self-promotion and hype and assume most here aren't either. At the same time, I view the various non-relational data stores as "competing" for mind share and I have to be pragmatic in terms of the tools I invest my time in. My fear is that 'laying low' in terms of the project's visibility will result in more people ignoring Voldemort and thus its eventual irrelevance.

I'll be frank -- I have a lot of questions that bother me:

1. Where are the non-LinkedIn core committers?
2. Where are the commercial "support and training" companies for Voldemort (a la Riptano, Northscale, 10gen, etc.)?
3. Where are the discussions about road map?
4. Where are the "success stories" of companies (outside of LinkedIn) using Voldemort -- and if there are any, why aren't they on the main web site?
5. Where are the stories about Voldemort in the media?
6. Where are the forthcoming books ("Voldemort - the Definitive Guide")?
7. Where are the add-on/plugin projects and vendors?
   

In my (unqualified) opinion - Voldemort is losing the battle in developer mind share. You're going to hate me for posting this, but this echoes the thought:

    http://regulargeek.com/2010/09/10/nosql-job-trends-september-2010/

I expect some hate mail here, but don't get me wrong -- I'm speaking from a position of commitment to the project. I want the Voldemort community to blossom past just LinkedIn and a few pilot projects.

What can I do to improve things? I've never participated in an open source community this much before, but I am willing to help. Here are some things I've thought about:

1. Refocusing my consulting company as a Voldemort professional services company
2. Start a local Meetup group for Voldemort committers, contributors, etc. (they have these for Hadoop and Hive, at least)
   
3. Speak at local user groups about Voldemort
4. Evangelize and find the success stories; turn on the marketing engine :)
   

My issue with the above is that I'm not sure there's interest enough (in the marketplace) to support these ideas. Also, I've never done any of the above :) But I'm willing to try.

What are your thoughts? Flames, answers, advice -- any input is welcome.

Thanks!
Kirk

[Jay Sprenkle]

I'm a newbie interested in trying out Voldemort.
I can tell you about my experience so far.

About me:
I started life as an electrical engineer and wanted to do computer design.
I morphed to programming over time for personal reasons.
I'm not a luddite but neither do I adopt technology for emotional reasons.
I use windows at work and Ubuntu at home.

My Experience:
I got interested in evaluating it because it's an almost perfect match for what I wanted.
I have a windows machine as do most corporate types. I can setup a Linux box to run
the server but wanted a C/C++ library to connect to it from windows. So far it seems there is
only code for linux C client. If I can't connect to the database from disparate environments
it's going to seriously limit it's adoption.

I couldn't find any help on this mailing list.

I've tried posting to the best technical support forum I've ever seen, stackoverflow.com.
On a forum where you can get most questions answered in minutes by very knowledgeable
people nobody has ever heard of the project. One poster answered my question about a windows port with "That's nice. I like pizza."

None of the provided source includes any names or email addresses to send questions to.

Suggestions to improve things:
If the project wants some adoption they need:

• Advertising/increased visibility to attract adopters
• Help for adopters if they run into problems

Once you get adopters the rest will take care of itself.

Solutions:

• Consider moving the mailing list to stackoverflow. It will get the project more visibility and will remove any requirement for mailing list maintenance. Take the forum web masters time and put him on answering questions about voldemort on that forum. (I have no association with SO, other than as a user).
  
• Ask one or more of the authors to spend some time helping new adopters.
• Offer a paid support option
• Start offering pre-built libraries for the most common environments.
• Create a position for a Voldemort advocate (perhaps the forum webmaster)
• Differentiate yourself from the other Voldemort
  

Good luck!

Jay

--
You received this message because you are subscribed to the Google Groups "project-voldemort" group.
To post to this group, send email to project-...@googlegroups.com.
To unsubscribe from this group, send email to project-voldem...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/project-voldemort?hl=en.

--
---
"Is real reporting dead?  News at 11!"

[Roshan Sumbaly]

Hi Kirk,

Thanks for bring up this point Kirk! Let me take this email a step ahead...

a) I'm starting a page on Github titled Roadmap feature requests. Lets make this the default page which the community should go to add their feature requests. In the past most features have been decided by the core committers without getting the community a chance to pitch in. I'll also put up our internal LinkedIn roadmap for Voldemort and we can get others to prioritize + modify it. 

b) "Powered by" page - Similar to Hadoop's PoweredBy page, another Github page . I would encourage users of Voldemort to start filling this up (along with links).  Once its filled up a bit, I'll personally shift it to our main site.

c) Meetups - To kick this off, I can definitely help in organizing the first meetup at LinkedIn itself. For starters we can have the LinkedIn team present how we use Voldemort internally. Or if anyone if willing to volunteer we'd love to hear how others use it.

Best,

Roshan 

--

[Roshan Sumbaly]

Also adding to this list a page dedicated to Presentations. I've seen multiple slides related to Voldemort flying around on slide share. This page should help in aggregating them in one place.

Best,

Roshan

[Geir Magnusson Jr.]

On Oct 6, 2010, at 6:17 PM, Kirk True wrote:

> Hi all,
>
> I'd like to start a non-technical discussion about Voldemort's community, visibility, and so on.
>
> First off, I love Voldemort. I'm a committer and I respect Jay, Alex, Roshan, Bhupesh immensely. I've been extremely blessed to work with those guys on Voldemort. I'm not into self-promotion and hype and assume most here aren't either. At the same time, I view the various non-relational data stores as "competing" for mind share and I have to be pragmatic in terms of the tools I invest my time in. My fear is that 'laying low' in terms of the project's visibility will result in more people ignoring Voldemort and thus its eventual irrelevance.
>
> I'll be frank -- I have a lot of questions that bother me:
>
> • Where are the non-LinkedIn core committers?

I'm one, but very inactive these days - mostly just due to time, my satisfaction with V - not lack of interest. My situation aside, this is an important point - committer diversity contributes greatly to the health and strength of a project.

> • Where are the commercial "support and training" companies for Voldemort (a la Riptano, Northscale, 10gen, etc.)?

It's a reasonable question, but... are they needed? I've been running V in our main, makes-all-the-money production systems at Gilt Groupe for over a year now and it's been great. I had one problem - due to a bug in a connection pool - and hitting the list or Jay/Alex directly solved it quickly.

> • Where are the discussions about road map?

Good question, but why *should* it keep evolving past a point? At some point, you hit the design target, or enough of it. What's missing now?

> • Where are the "success stories" of companies (outside of LinkedIn) using Voldemort -- and if there are any, why aren't they on the main web site?

I'd be proud to post ours.

> • Where are the stories about Voldemort in the media?

> • Where are the forthcoming books ("Voldemort - the Definitive Guide")?
> • Where are the add-on/plugin projects and vendors?
>
> In my (unqualified) opinion - Voldemort is losing the battle in developer mind share. You're going to hate me for posting this, but this echoes the thought:
>
> http://regulargeek.com/2010/09/10/nosql-job-trends-september-2010/

I don't understand what it means to be a 'nosql job'. Does that mean the job is focused on the named tech, or that the named tech is involved?

If the latter, why isn't memcached there? That's arguably a 'nosql technology' and I imagine is many, many orders of magnitude higher. (Every Rails job is a 'nosql job' then, right? :)

>
> I expect some hate mail here, but don't get me wrong -- I'm speaking from a position of commitment to the project. I want the Voldemort community to blossom past just LinkedIn and a few pilot projects.

We're not "pilot". We're "took us from $90MM to $500MM and we're still growing..."

>
> What can I do to improve things? I've never participated in an open source community this much before, but I am willing to help. Here are some things I've thought about:
>
> • Refocusing my consulting company as a Voldemort professional services company

That would be great, but is there demand?

Here's my concerns :

1) I've been involved in OSS in one way or another for the last decade, and my preference is a vibrant, supportive community here on the list (a la apache projects of yore). That's a strawman and I'm sure not what you are imagining, but I thought I'd throw it [clumsily] out there.

2) You're a valuable asset in this community. do you worry about being in conflict? Would you put as much energy into helping people on the lists when your fiduciary duty to yourself/family is to generate revenue.

> • Start a local Meetup group for Voldemort committers, contributors, etc. (they have these for Hadoop and Hive, at least)

The latter are harder to use :)

I volunteer to try this in NYC.

> • Speak at local user groups about Voldemort

Happy to do that as well. I've given a few talks (web2.0, qcon, ETE, etc) about V, and always happy to do more.

> • Evangelize and find the success stories; turn on the marketing engine :)

I'm tired :)

>
> My issue with the above is that I'm not sure there's interest enough (in the marketplace) to support these ideas. Also, I've never done any of the above :) But I'm willing to try.
>
> What are your thoughts? Flames, answers, advice -- any input is welcome.

I'm a satisfied user - my life depends on voldemort. I talk about it. I recommend it. But the damn thing just works. ALso, maybe it's too niche? I guess there's a set of design patterns around using V - I'd be happy to share ours - but it's just not the same as MongoDB or Couch in terms of being able to scale in the small (back your PHP-based blog), and IIRC, there haven't been any high-profile disasters with the tech (e.g. Cassandra at Digg...)

One thing we might do is try to get some tools and related going. At Gilt, we wrap V in a 'service' that only speaks JSON, so clients really don't know it's V or have to speak V. They have to deal with the API semantics of versioning, but it's not a big deal - we've adapted our ROR codebase via ActiveResource to work with it just fine.

I'm happy to get our so-called "kvstore" out in OSS - maybe others have similar things they've developed?

geir

[Mark Rambacher]

To add the discussion, last week we put in place our first production
Voldemort cluster. This cluster will contains 20+ nodes and hold over
5 TB of master (non-replicated) data . We plan to make extensive use
of Voldemort in the future -- this is just the first of many potential
installs.

I am a Voldemort devotee. I think the technology is great and
generally works very well. The pluggable design of the system and
extensibility makes it ideal for us. Having said that, I fear that
going forward we will loose the battle or separate from the community
if the status quo continues. I see Voldemort using the mindshare
battle. I have to (fairly regularly) justify why we picked Voldemort
over other technologies, both to our management and our customers.
Better press for Voldemort would certainly make that job easier. Even
the chatter on this group has dropped off substantially over the past
few months.

We have attempted to be good open-source citizens and have contributed
changes and bug fixes back to the Voldemort community. Unfortunately,
none of our changes have made it back into the mainline yet. In the
year or so we have been using Voldemort, we have seen several new
features developed (Slop Store, Rebalancing, Zoning, AdminClient).
Most of these were developed without much input or vetting from the
community (the Wiki links were done after an implementation and not
before, making it harder to change directions or accept feedback). As
currently coded, we typically find the implementations do not meet our
use cases directly and require further development on our part. We
are not adverse to doing or helping with the work, it would just be
easier to do it up-front rather than try to retrofit everything in.

I do not see Voldemort as being "fully implemented" yet. For example,
a way of sending changes to a secondary store (such as for backups)
would be a useful feature. An asynchronous communication protocol
would be nice. Other routing strategies (such as "startsWith"
routing) and the additional store methods (listKeys) are also
necessary. The list of features goes on and on -- better
configuration management, performance and resource optimizations,
"read your own writes" consistencies, column-oriented storage, etc.

I think the Voldemort development community needs more process around
it:
- How and when bug fixes are reviewed and committed back to the
mainline and delivered in subsequent release needs to be published and
followed;
- A process needs to be documented and followed for "minor features"
-- enhancements that are little more than bug fixes but not as in-
depth as a major design feature.
- When new "major features" are being added, the new design should be
published the design to the community before it implemented. At this
point, the community can make recommendation on changes to the design
(Are wikis or the mail list or something else the best way of holding
this review?). What ends up getting implemented should depend on the
feedback of the community, the disruption to current installations,
the future extensibility and pluggability of the feature, and the
level of commitment (pigs, chickens, etc) from other community
members.
I do not think we need the full "Apache" process, but something needs
to be more formally codified and followed. Even the release schedule
(it was monthly in the spring but now is???) needs to be more
formalized and published.

I believe better process and more publicity will keep Voldemort
vibrant and will lead to more companies using it and more success
stories down the road. Voldemort needs to publicize these success
stories (Cassandra has multiple users listed on its front page) in
order for it to continue to flourish.

- Mark

[Maarten Koopmans]

Well, IMHO what Apache brings to the table is:

- a well defined legal framework (licenses)

- a well defined process from incubator, to voting for releases, code contributors etc.

- and visibility

But even then, projects fail/go past their due date/....

Putting those practices in place isn't that hard, but it make it soooo much easier if you want a community.

If you just released your amazingly piece of software (no cynicism here) as open source, fine too.

So what it boils down to: does the core team want a vibrant community, or release their work as open source 

and do not want to be bothered by running a community as well (which is *very* time consuming).

Either way, I'm happy with Voldemort. And I agree with the earlier observation of hitting your design goal. After

that it should be just maintenance.

--Maarten

[Kirk True]

Hi Geir,

Thanks for the feedback. Questions/comments/etc. inline...

>> � Where are the commercial "support and training" companies for Voldemort (a la Riptano, Northscale, 10gen, etc.)?

> It's a reasonable question, but... are they needed? I've been running V in our main, makes-all-the-money production systems at Gilt Groupe for over a year now and it's been great. I had one problem - due to a bug in a connection pool - and hitting the list or Jay/Alex directly solved it quickly.
>

It's one of those things that ideally wouldn't be needed, but my
question is - is the lack of such companies indicative of lack of belief
in the project or--speaking in the positive--a testament to the quality
of answers on the ML and IRC?

>
>> � Where are the discussions about road map?

> Good question, but why *should* it keep evolving past a point? At some point, you hit the design target, or enough of it. What's missing now?

True, we shouldn't add things just for the sake of it. And if there
isn't anything missing, then that should be stated. But I know that the
guys at LinkedIn have to work on something, right? :)

>> � Where are the "success stories" of companies (outside of LinkedIn) using Voldemort -- and if there are any, why aren't they on the main web site?

> I'd be proud to post ours.

If you could post it to Roshan's newly added page
(http://github.com/voldemort/voldemort/wiki/Powered-By) that would be
awesome!

>> � Where are the stories about Voldemort in the media?
>> � Where are the forthcoming books ("Voldemort - the Definitive Guide")?
>> � Where are the add-on/plugin projects and vendors?

>>
>> In my (unqualified) opinion - Voldemort is losing the battle in developer mind share. You're going to hate me for posting this, but this echoes the thought:
>>
>> http://regulargeek.com/2010/09/10/nosql-job-trends-september-2010/
> I don't understand what it means to be a 'nosql job'. Does that mean the job is focused on the named tech, or that the named tech is involved?
>
> If the latter, why isn't memcached there? That's arguably a 'nosql technology' and I imagine is many, many orders of magnitude higher. (Every Rails job is a 'nosql job' then, right? :)
>

Sure - I question the metrics too :)

But it *is* a data point that engineering managers, developers, etc. use
to determine interest, maturity, community, availability of talent, etc.

>> I expect some hate mail here, but don't get me wrong -- I'm speaking from a position of commitment to the project. I want the Voldemort community to blossom past just LinkedIn and a few pilot projects.
> We're not "pilot". We're "took us from $90MM to $500MM and we're still growing..."

Great. We need more of these :)

>> What can I do to improve things? I've never participated in an open source community this much before, but I am willing to help. Here are some things I've thought about:
>>

>> � Refocusing my consulting company as a Voldemort professional services company

> That would be great, but is there demand?
>
> Here's my concerns :
>
> 1) I've been involved in OSS in one way or another for the last decade, and my preference is a vibrant, supportive community here on the list (a la apache projects of yore). That's a strawman and I'm sure not what you are imagining, but I thought I'd throw it [clumsily] out there.
>
> 2) You're a valuable asset in this community. do you worry about being in conflict? Would you put as much energy into helping people on the lists when your fiduciary duty to yourself/family is to generate revenue.
>

Yes, I don't really want to take this approach :)

But the points you make are correct - my personal aim is to make the
community stronger. I believe the other things will fall naturally into
place (for everyone).

>> � Start a local Meetup group for Voldemort committers, contributors, etc. (they have these for Hadoop and Hive, at least)

> The latter are harder to use :)
>
> I volunteer to try this in NYC.
>

Roshan mentioned trying to start something up here in the valley. Do we
know where the core committers and other interested parties are located,
geographically-speaking?

> I'm a satisfied user - my life depends on voldemort. I talk about it. I recommend it. But the damn thing just works. ALso, maybe it's too niche? I guess there's a set of design patterns around using V - I'd be happy to share ours - but it's just not the same as MongoDB or Couch in terms of being able to scale in the small (back your PHP-based blog), and IIRC, there haven't been any high-profile disasters with the tech (e.g. Cassandra at Digg...)

To be fair, was Cassandra ever outed as the culprit?

> One thing we might do is try to get some tools and related going. At Gilt, we wrap V in a 'service' that only speaks JSON, so clients really don't know it's V or have to speak V. They have to deal with the API semantics of versioning, but it's not a big deal - we've adapted our ROR codebase via ActiveResource to work with it just fine.

Isn't there a JSON frontend being discussed somewhere?

Thanks guys!

Kirk

[Mingfai]

On Thu, Oct 7, 2010 at 10:27 PM, Mark Rambacher <mram...@gmail.com> wrote:

To add the discussion, last week we put in place our first production
Voldemort cluster.  This cluster will contains 20+ nodes and hold over
5 TB of master (non-replicated) data .  We plan to make extensive use
of Voldemort in the future -- this is just the first of many potential
installs.

Thanks for this thread and Mark's sharing of his successful story. I'm new to Voldemort and the information are very useful to me.

I want to add 2 points:

• if Voldemart is available in a Maven repository, it will help people to start trying it. As I understand, there was attempt(s) or at least discussion about this and it seems the project is not structured in a way to allow building a client jar easily.
• The repository in github doesn't seem to use any tag at all. Some people may concern about the release management process of a project in evaluation.
  

Regards,
mingfai

[Kirk True]

I would volunteer to do a Maven client POM. I've never pushed to the central repos, but I'm willing to make a go for it. :)

Kirk

Sent from my iPhone

--

[Tatu Saloranta]

On Sat, Oct 9, 2010 at 3:41 PM, Kirk True <ki...@mustardgrain.com> wrote:
> I would volunteer to do a Maven client POM. I've never pushed to the central
> repos, but I'm willing to make a go for it. :)

If necessary I can help with getting Maven artifacts deployed via
Sonatype repository. Since github does not offer direct Maven sync
service (as far as I know), Sonatype's free OSS project hosting (see
https://docs.sonatype.org/display/Repository/Sonatype+OSS+Maven+Repository+Usage+Guide)
is a good way to publish Maven artifacts to central Maven
repositories. This is quite easy to do, but obviously some work first
time; although not a ton after reading that article and cut'n pasting
setup from one of many existing projects that use it.

-+ Tatu +-

[Maarten Koopmans]

But, please, let's use Maven as an option besides the current distro. I for one am not too happy with ant/Maven/SBT etc. (no flames intended).

[Tatu Saloranta]

On Sun, Oct 10, 2010 at 3:05 AM, Maarten Koopmans
<maarten....@gmail.com> wrote:
> But, please, let's use Maven as an option besides the current distro. I for
> one am not too happy with ant/Maven/SBT etc. (no flames intended).

I think there are multiple parts to this:

(a) Exposing artifacts via Maven. I think this is a must for Java OSS
projects by now
(b) Building project using Maven. This is pretty much optional.

There are ways to achieve (a) from other build systems (as well as
using Maven repositories for external parts). I have projects with Ant
builds that use maven-ant-tasks for publishing jars to Maven repos for
example.
It is often just as easy to do both, which is why complete conversions
are common.

But I guess what you are saying is that you would not want to be
forced to use Maven for downloading distribution? I think that can be
done even if using Maven as build system.

-+ Tatu +-

[Mingfai]

On Mon, Oct 11, 2010 at 12:12 AM, Tatu Saloranta <tsalo...@gmail.com> wrote:

On Sun, Oct 10, 2010 at 3:05 AM, Maarten Koopmans
<maarten....@gmail.com> wrote:
> But, please, let's use Maven as an option besides the current distro. I for
> one am not too happy with ant/Maven/SBT etc. (no flames intended).

I think there are multiple parts to this:

(a) Exposing artifacts via Maven. I think this is a must for Java OSS
projects by now
(b) Building project using Maven. This is pretty much optional.

I'm actually able to build the whole project and install to my local repository without changing the current directory structure. Only one of the compile dependency, that is Thrift, is not available in central. However, around 1/3 of the tests fail.

I personally like Maven as it allows managing dependency easily, encourage a a module structure, and separates different kinds of src clearly. AFAIK, Ant with Ivy should be able to handle the dependency part but I haven't used Ant for a long time.

Anyway, it seems most people mainly care about a client jar. And it probably require restructuring the project to create a client module,  or any unusual way to build (for maven, I guess it is possible to use build-helper-maven-plugin to select the specific packages)


regards,
mingfai
 

[Alex Feinberg]

A build/tools engineer has actually volunteered to help convert our
current build system to use ivy, which will make it very easy to
export a .pom xml (as well as make it easy to keep up with latest
version of the jars we rely on for Voldemort). I'll follow up with him
during the week. I think making the voldemort jar + pom available via
maven repo is quite important as well.

- Alex

On Sun, Oct 10, 2010 at 9:12 AM, Tatu Saloranta <tsalo...@gmail.com> wrote:
> I think there are multiple parts to this:
>
> (a) Exposing artifacts via Maven. I think this is a must for Java OSS
> projects by now
> (b) Building project using Maven. This is pretty much optional.
>
> There are ways to achieve (a) from other build systems (as well as
> using Maven repositories for external parts). I have projects with Ant
> builds that use maven-ant-tasks for publishing jars to Maven repos for
> example.
> It is often just as easy to do both, which is why complete conversions
> are common.
>
> But I guess what you are saying is that you would not want to be
> forced to use Maven for downloading distribution? I think that can be
> done even if using Maven as build system.
>
> -+ Tatu +-
>

[Kirk True]

Hi Mingfai,

> Anyway, it seems most people mainly care about a client jar. And it
> probably require restructuring the project to create a client module,
> or any unusual way to build (for maven, I guess it is possible to use
> build-helper-maven-plugin to select the specific packages)

I would assume that building a true *client* JAR would require some
code-level changes to move things around and separate things along that
boundary.

Thanks,
Kirk

[Maarten Koopmans]

But I guess what you are saying is that you would not want to be
forced to use Maven for downloading distribution? I think that can be
done even if using Maven as build system.

That's exactly what I am saying! Just give me one big fat jar (with all dependancies included as well).

--Maarten 

[Kirk True]

Hi Alex,

What's the status on getting the build system to spit out a POM? I'm
starting a new project soon and would love to use Ivy/Maven (/me ducks).

Thanks,
Kirk

[Maarten Koopmans]

And SBT for Scala is then also an option. Plus I have a pretty decent
Scala client wrapper for those who care in the community.

[Jferg]

On the topic visibility - small things? We had a NoSQL talk at work -
Wikipedia - had no entry for Voldemort. I saw Couch and I think Mongo
DB. So when people go look up NoSQL... what do they see?



On Nov 13, 9:48 am, Maarten Koopmans <maarten.koopm...@gmail.com>
wrote:

> And SBT for Scala is then also an option. Plus I have a pretty decent
> Scala client wrapper for those who care in the community.
>

> On Saturday, November 13, 2010, Kirk True <k...@mustardgrain.com> wrote:
> > Hi Alex,
>
> > What's the status on getting the build system to spit out a POM? I'm starting a new project soon and would love to use Ivy/Maven (/me ducks).
>
> > Thanks,
> > Kirk
>
> > On 10/10/10 2:04 PM, Alex Feinberg wrote:
>
> > A build/tools engineer has actually volunteered to help convert our
> > current build system to use ivy, which will make it very easy to
> > export a .pom xml (as well as make it easy to keep up with latest
> > version of the jars we rely on for Voldemort). I'll follow up with him
> > during the week. I think making the voldemort jar + pom available via
> > maven repo is quite important as well.
>
> > - Alex
>

> > On Sun, Oct 10, 2010 at 9:12 AM, Tatu Saloranta<tsalora...@gmail.com>  wrote:
>
> > I think there are multiple parts to this:
>
> > (a) Exposing artifacts via Maven. I think this is a must for Java OSS
> > projects by now
> > (b) Building project using Maven. This is pretty much optional.
>
> > There are ways to achieve (a) from other build systems (as well as
> > using Maven repositories for external parts). I have projects with Ant
> > builds that use maven-ant-tasks for publishing jars to Maven repos for
> > example.
> > It is often just as easy to do both, which is why complete conversions
> > are common.
>
> > But I guess what you are saying is that you would not want to be
> > forced to use Maven for downloading distribution? I think that can be
> > done even if using Maven as build system.
>
> > -+ Tatu +-
>
> > --
> > You received this message because you are subscribed to the Google Groups "project-voldemort" group.
> > To post to this group, send email to project-...@googlegroups.com.
> > To unsubscribe from this group, send email to project-voldem...@googlegroups.com.

> > For more options, visit this group athttp://groups.google.com/group/project-voldemort?hl=en.

[Jferg]

Okay it got a mention under a sub-link, but eh... not the front page.
I suggest doing lightening talks at work about it. I did that. Boy
did it get some reactions - from the SQL DBAs. I narrowly avoided a
grand inquisition. However, it had he effect I wanted. People either
attended or watched my recording later. We now have a NoSQL lunch
meeting and when we discuss principles there I will add my 2 cents
with how Voldemort addresses it. I also have made at least 8 people
(including my lead Architect) aware of Voldemort's existence and
features.

My advice do lightening talks at work about it, get the geeks
listening and attract the attention of SQL DBAs too.