[xwiki-devs] Back to the future of XWiki !

Denis Gervalle dgl at softec.lu
Fri Mar 4 15:05:44 UTC 2011

On Fri, Mar 4, 2011 at 14:58, Vincent Massol <vincent at massol.net> wrote:

> Just a quick mail to say that I agree a lot with what Andreas said in his
> mail: I also think we need to do tech marketing a lot more (but it's
> everyone's duty to do so, not specifically Ludovic's):

I agree fully.

* have xwiki talks at more conferences (by xwiki devs and by users)
> * have more blogs about xwiki out there
> * have more articles on xwiki
> * have books on xwiki ;)

I have already think about that, but development is moving so quickly that
this is a big change to covers all recents features...

> * have xwiki installs at more known forges (we're already on ow2): apache,
> codehaus, etc
> * have xwiki in standard linux distributions
> * move to github for easier forking/contributions
> * push xwiki artifacts to central
> * make xwiki more easily reusable by other projects to spread the brand:
> that's what we're doing with XWiki Rendering and XWiki Commons
> I also agree that I'd prefer to see way more stabilization of the xwiki
> platform (polishing, bug fixes, improvements of existing features) rather
> than add lots of new features.

My very big +1 on this point.

Currently, the most important feedback from our customers says that they
feel the product to be too much complex and are afraid by having so much
choice in the interface. We have try to mitigate that by adding our own
"interface hiding" feature, trying to remove as much as possible, and not
show what you have no rights for or is not active. As you can imagine, this
is the opposite of adding features.

To be even more concrete, let me take the last request I have received. It
was about having a single "Connexion" links in place of having "Register"
and "Identification". My customer was saying: "how can you live without what
all other sites have, a single login form that allows to register when you
do not have an account ?" I know changing this is not so difficult, but it
should be built-in IMO.
You can't imagine how much support we do just about login, lost password,

As another example, I was trying to made a very minimalistic interface for a
very small group of non-technicians that would like to share some ideas.
Having a full wiki just for them is overkill, so I made a space. But without
changing .vm, its is almost not possible to have a simple interface (even
with our hiding feature). I hate changing vm since it is a pain to maintain
the changes over releases. After an hour of work, I finally just left of due
to the top left search box that was not working properly (it was a 2.4, I
had to fix a {html} issue, of course I know some changes may have been done
since then, but anyway...). Basically, when you search in the search box,
you leave the current space, which may cause skin changes (terrible for an
end-user accessing only a single space), and search results shows  a listbox
of all spaces, while you may only access a few. We had finally use Google
Docs :(

I regrets not to have much time now to works on these simple interface
issues that could be a real barrier for basic users. I really encourage any
developments that improve remaining legacy code or that contribute to
simplifying or clarifying the interface. Cathy do a great job at this, but
implementation does not always follows. If I would have time, I would have
focus my time on simplifying, and having an interface for basic and advanced
users more separated, allowing to show features only if requested.

XWiki is now a big piece of software, we should be careful not to afraid
users with overwhelming features, and ensure that basic and already existing
features are fully featured but simple.

> But I know Ludovic is aware of that and it's our collective duty (everyone
> here: devs, users, contributors) to balance stabilization vs new features.

Of course, I agree that I am not being of much help currently :(


> Thanks
> -Vincent
> On Mar 4, 2011, at 2:27 PM, Jerome Velociter wrote:
> > Hi Andreas, everyone
> >
> > On Fri, Mar 4, 2011 at 10:48 AM, Andreas Hahn <ahahn at gmx.net> wrote:
> >> Hi,
> >>
> >> I'm a bit surprised that you are just talking about features and not
> >> about moving from a niche product to get more mainstream.
> >> When I'm saying niche product I'm just referring to what google trends
> says.
> >> http://www.google.com/trends?q=xwiki%2C+foswiki%2C+twiki%2C+confluence
> >
> > "confluence" is a word of the English language, it makes this chart off
> by far.
> >
> > This one would be a bit more relevant in my opinion :
> > http://www.google.com/trends?q=xwiki%2C+atlassian%2C+mindtouch& but I
> > still don't think it means that much.
> >
> >>
> >> Well I think that almost all people on this mailing list like xwiki a
> >> lot and the hard and devoted work the developers are doing and all wish
> >> xwiki and its backing company all the best and a long and prosperous
> future.
> >>
> >> However I'm concerned when reading that the XWiki SAS 'Product
> >> Marketing' lead (and CEO) thinks that it helps the product to get more
> >> involved as a developer.
> >> IMHO a product marketing lead should have totally different priorities.
> >
> > Ludovic will probably answer this for himself, but as a quick remark I
> > think you've misinterpreted his mail. He doesn't say he will come back
> > as a developer, but only as a Product Marketing lead, thus working
> > close to developers and the community, because, well, that's how
> > (real) open source software is developed.
> >
> >>
> >> IMHO its a misconception to think that more features will help the
> >> product to attract a greater audience.
> >>
> >> Just a few things that come to my mind:
> >>
> >> * Push xwiki into as many linux distros as possible with 'one click
> >> installers'
> >> * Create ready-to-go images for popular cloud hosters (such as Amazon
> EC2)
> >> * Talk to as much hosters and push them to introduce XWiki as part of
> >> their offerings.
> >> * Make it simple and attractive for people currently using other
> >> software (Wordpress comes to my mind) to move to the more powerful XWiki
> >
> > I think this vision is not complete.
> > Cloud and hosted software got a lot of attention lately, but in-house
> > software is not dead yet. Maybe in a couple of years in-house will
> > mean "deployed on a private cloud", but we are still far from here, we
> > would have first to see a complete, mature, open source cloud solution
> > emerge, and have organizations deploy it on their infrastructures. I
> > think we are speaking of half a decade here, so we have time to try
> > and not miss that train :)
> > In the mean time, I think XWiki's greatest challenge is to become the
> > de-facto standard in the enterprise open source stack when speaking
> > collaboration. Right now we are not badly positioned when speaking of
> > "collaborative edition/redaction". We need to affirm this position and
> > push it further to "collaboration" in a larger meaning (without trying
> > to be the solution to everything or a social network tool, because
> > that's not what XWiki is).
> >
> > Now I agree with you, it would be of course a good thing to have the
> > EC2 distribution, the VMWare images and be present in more online
> > offerings etc. Just I think it's not absolutely critical versus making
> > XWiki a product so great that it just becomes a logical choice for
> > collaboration.
> >
> > I also agree having easy installs on linux would be great to increase
> > visibility. But as a I understand, the cost of packing java apps for
> > linux is quite high. This is why you don't see packages even for
> > popular java (web) applications.
> > Trying to replace all blogs out there by XWiki is not going to help
> > too much IMHO. Even if we geeks like it to use our toy for absolutely
> > everything, for the rest of the world, XWiki remains mainly a
> > professional software for doing collaborative edition.
> >
> > That was my 2 cents. Not trying to just "defend my turf" as one of the
> > developer of XWiki, but rather trying to balance the vision that says
> > we need to be very strong on the web / commercial clouds to win. There
> > already have been some success stories of open source end-user
> > applications (in the sense:  not infrastructure apps) a la Alfresco,
> > Liferay, Zimbra, etc. I think XWiki is a candidate for such, even if
> > we are not there yet. The important part in my opinion is building a
> > great product, the rest will come naturally. That's why having strong
> > product marketing with Ludovic being active on features investigation
> > and design make sense and will be a good thing in my opinion.
> >
> > Jerome.
> >
> >>
> >> I may be paranoid but I wonder if we will see XWiki survive the next
> >> couple of years if its not gaining momentum in the big world.
> >>
> >> just my 2 cents
> >>
> >> Andreas
> >>
> >>
> >>
> >>
> >>
> >>
> >> Am 03.03.2011 14:21, schrieb Ludovic Dubost:
> >>>
> >>> Hi all,
> >>>
> >>> As many of you may know, 7 years ago, I created the XWiki Open-Source
> >>> software.  A few years ago, especially when Vincent arrived, I took a
> >>> step back from development of the XWiki product to focus on
> >>> developping XWiki SAS which allows to support the XWiki development.
> >>> I've particularly spent my time making sure that the deployments of
> >>> XWiki our Customers have been doing are successfull. I've left the
> >>> product development work to committers who under the lead of Vincent
> >>> have done amazing work in the last few years. We have provided some
> >>> product marketing in addition through Guillaume's and lately Gregory's
> >>> work which allowed to bring some feedback from users and customers and
> >>> also bring a different less technical perspective to the XWiki Product
> >>> development. Cati also joined the team and allowed us to make huge
> >>> steps forward in product usability and design.
> >>>
> >>> Now, as XWiki SAS's project implementation team is doing great work on
> >>> it's own, I've decided to involve myself more in the future of XWiki's
> >>> product. I'm now the new XWiki SAS 'Product Marketing' head which
> >>> means I will involve myself in community discussions about the product
> >>> features actively developped, investigated for future development as
> >>> well as in general discussions about the XWiki Software's future. I
> >>> will try to bring the knowledge that XWiki SAS's customer, project
> >>> managers, support team bring us from using XWiki in production
> >>> environments to the community so that we make better decisions, more
> >>> focused on developments that will allow to increase XWiki's success
> >>> with end users. At XWiki SAS we have already done a lot of work to
> >>> organize this feedback so that we know more things that XWiki SAS's
> >>> developers and the XWiki community should work on.
> >>>
> >>> In the next few weeks you will therefore see me work on the
> >>> investigations of future features as well as work more closely with
> >>> developers whenever they feel the need on the features they implement
> >>> actively. I will also manage the different surveys we have done in the
> >>> past to gather feedback from our community. These surveys have still
> >>> been running on the xwiki.org web site although they are now quite
> >>> outdated. I will propose to review them and launch them again.
> >>>
> >>> If users or developers have any feedback about XWiki, about what we
> >>> should work on in priority, please do send that feedback on the list
> >>> (preferably). Although we have a much bigger list of great ideas than
> >>> what we currently can achieve, you can help us on helping us select
> >>> the most important ones that will make a difference.
> >>>
> >>> I would like to use this occasion to thank our great community of
> >>> developers and users who help the committers every day to make XWiki
> >>> better. We could not do this without all the suggestions, ideas,
> >>> patches that our community provides.
> >>>
> >>> Thanks
> >>> Ludovic
> _______________________________________________
> devs mailing list
> devs at xwiki.org
> http://lists.xwiki.org/mailman/listinfo/devs

Denis Gervalle
eGuilde sarl - CTO

More information about the devs mailing list