I realized TimDistro is much bigger can of worms If you are in hurry, you can go directly to #WhyNewWeb and vote for (or against) in PleaseCreateNewWeb.

It looks to me there are three kinds of developers interested in Twiki.

One is wildly successfull installations like those of PeterThoeny and JohnTalintyre, (and other CoreTeam memebers), with mission-critical Twiki deployment with many webs and thousands of pages succesfully deployed, hundreds of users trained on current markup language and ways to do things, complex workflows of processes in production relying on current functionality of plugins. For these lucky folks, 100% backward compatibility of any enhancement is essential, and rightly so.

But not every member of Twiki community is in such an enviable position. Many times, Twiki deployment is only moderate success with only handfull of users (editors), couple dozens of pages or less, and it is not mission critical at all. Or somebody just likes feature set and tries to play with Twiki. Or my case: Twiki instal is almost a flop with one or two reluctant users which occasionally will add some content is asked to do so, but I am alone to enjoy excitement of any new feature, and I even do not care to make some basic Twiki features run (like WebNotify).

For all these "hobby" Twikis, 100% backward compatibility is non-issue. Heck, I can manually edit and resave all pages on our installation within a day - substantially shorter time than what I already spent to upgrade from Athens to Beijing. And I personaly know more than one user who downloaded Twiki and was not able to make it run, so switched for another simper wiki (and is sorry for missing features). I really appreciate all work done in adding security and performance of Twiki, but what I need is increased ease of install for me and usability for my users.

Then, there is third group of (AFAICT perl gurus) developers who like current Twiki features and would like to play with it and extend it. Of course changes will be disruptive for a (maybe long) while, but "hobbyists" do not care too much.

There were accusations that any changes not sanctified by 100% bacwkard compatibility are by definition a fork, and people asking for it are close to heretics (of course I am exagerating ). I prefer to see it as if a group of scouts runs ahead of main army to look what dangers are ahead, try multiple different ways how to handle it, and when ready, join the best method back to main trunk for everybody to enjoy. What is bad about it? Then, these scouts want to pack changes conveniently so "hobyists" can easier to try them. Why not?

So what I wanted to initiate was a discussion about next generation of Twiki. Obviously, when MainPeterThoeny started with Twiki couple of years ago he could aim only for doable goals, and wanted results quickly. We have luxury not to hurry (we have running stable Twiki, ThankYouPeterThoeny). We know what issues could arise when introducing Twiki to new users, what if using Twiki in production, what is important for users, and have much more developers with multitude of different skills on board (I strongly hope ), so our goals and approaches might be different. What we need now is a meeting place where we can all think what coud be a better way to do stuff, what are limitations of our current approach, and how to improve it.

I was thinking about creating new web for discussion about Twiki Next Generation. (I need a provisional name - let's say Tng web to make Star Trek fans happy ). I believe (and might be wrong - I was before) that Codev web is focused on evolutionary changes, not compromising stability (and it's fine with me). While discussion in Tng web will certainly use many smart ideas from Codev, Next Generation in Tng will not be burdened by compatibility. Simple to use upgrade converter should suffice. New Tng web will remove the noise from Codev web for those not interested. Also, separate web will allow us to use different skin - which is IMHO important, because I hope to invite external experts on particular issues and do not want them be overhelmed (or pushed away) by skin in Codev web. And if these experts will wander a little in Tng web, I do not want them to get lost in layers of Codev discursions (which many times might be obsolete in Tng).

I am definitely not a person who can make Next Generation happen alone. I am not Perl guru. I need help. But IMHO there are some developers interested in this kind of project, but maybe reluctant to ask for it. I hope these developers will join me to discuss what might be next generation of Twiki (and might be not, if not enough people interested in coding will join).

So I am not a benevolent dictator - I am agent of change . I have only one goal: make Twiki easier to install, administer, understand, use, and upgrade. Backward compatibility is fine but not important.

I have no idea how (and if) it is going to happen. But let's talk what it is all about, starting with project name and features.

PeterThoeny warned that the more noise somebody (like me) raises the less is he interested to answer. So I think he might not read this page at all, and I created another page, PleaseCreateNewWeb. Please go there and vote there for (or against) new web. I hope he will read Please... page and decide. You can discuss anything for or against on this page, and put just your short comments in PleaseCreateNewWeb page - to make it easier to read. Or not, do as you feel is right.

So all I wanted for TimDistro before are just minor details now, but I preserved my original list of improvements here for history.

• Create docs for newbie users only
• Rename "Main" web to "User" and make pages writeable only by user
• Create another "Home" web with writeable user homepages
• rename "Topic" to "page", and make "web" a parameter (with pre-installed value of "zone")
• Install bunch of stable skins, ready to be enabled
• add bunch of stable plugins, disabled by default

Of course we will talk about it (I hope) in Tng web. Or here in Code. Or not at all. We will see.

I know I need to refactor TimDistroDev. Where the time goes?

-- PeterMasiar - 01 Jun 2003