Audience:
Tiki has a development model which is different than most of the other comparable projects. (really!) In a nutshell: Software made the wiki way and the FLOSS Web Application with the most built-in features.
The "Tiki model" consists of:
Marc Laporte, a long-time Tiki Administrator and Tiki Community Software Association board member are interviewed for the 146th FLOSS Weekly netcast, with Randal Schwartz and Simon Phipps, current President of the Open Source Initiative (OSI)
It's a Do-ocracry. Everyone is a volunteer. Things get done because someone like you decides to take the time to make it better. Join a team!
Participate in a TikiFest!
We use this method for commits to the stable branch (Ex.: 3.1 -> 3.2 -> 3.3) but when developing the new major version (ex.: 3.x to 4.0), all contributors commit directly to the main code base. We feel this is a great balance of
* having lots of innovation and low coordination overhead when developing the future version.
* while having minor versions with as few regressions as possible.
So if you need a lot of stability you should skip .0 releases (which really, you should avoid for any project) and have an upgrade pattern of 3.1 -> 3.2 -> 4.1 -> 4.2 -> 5.1, etc
Scheduled releases are a fundamental part of the Tiki model. As a large community, it makes it clear for everyone what to expect.
There are two major releases per year (October and April), with several minor releases in between. Please see: Roadmap and Version lifecycle and When to release - think popcorn
Whatever is not ready in time for a release is pushed back to the next. If you are late for a train, you may miss it, but you'll be early for the next one ๐
4 was too short. 12 was too long.
6-8 months was just about right. It's long enough to make big changes, yet it's short enough that you don't have to wait forever. For people in a business setting, it's essential for them to be able to a commit to a date with their customer. By having a set date, there is an incentive to commit to trunk, and know when it will become stable. 6 months has the benefit that we can choose good months to release (October and April) that are the same every year. In the future, it may become even faster (3-4 months), but it will certainly not become longer than 6 months. People who need a slower cycle can use LTS.
During a 6 month period, there are several thousand commits, which is more than enough to justify a major release. Just check Tiki2, Tiki3, Tiki4, Tiki5, Tiki6, Tiki7, Tiki8, Tiki9, Tiki10, Tiki11, Tiki12, Tiki13 and RoadMap
All functionality is in the core. Developers are collaborating throughout the whole co
Releases take more planning and coordination because all the features are released at once, but since it's de facto Synchronized Releases, the whole community can collaborate right away on the next release and this permits rapid major releases (even with such a vast code base). See lifecycle.
Tiki is a "1-stop shop" with tons of features. In fact, Tiki is the FLOSS Web Application with the most built-in features. However, there is but one forum, one blog, etc. Having duplicate features would lead to all kinds of problems and would splinter our user & developer base. If several people want to improve a feature, they collaborate and extend, in the wiki spirit. This is why we can give highly descriptive names like Tiki blog, Tiki forum, etc ๐ All features are optional.
On Wikipedia, when two pages are about the same topics, they are merged:
http://en.wikipedia.org/wiki/Wikipedia:Merging
In Tiki, it's the same idea but for features.
We eat our-ou own Dogfood and we hope you will as well ๐ All *.tiki.org sites are powered by unmodified versions of Tiki. And new versions of Tiki power the sites before the official versions are released.
Tiki is a do-ocracry. Join a team!
The vast majority of decisions go to the ones actually doing the work (be it code, translations, documentation, etc). There is a very strong collaboration culture in Tiki. So contributors discuss and find better ways to do things. You may have the impression that this is a big and complicated thing, but it's not. The vast vast majority of things to decide, everyone agrees. (it's easy to agree on the ideal; the real challenge is deciding on what you can do with limited resources, but since it's a do-ocracy, this solves itself.)
The project admins as listed on WhoWhat. For example, they act as a safe guard, as a last resort to deal with Difficult People.
Anyone can take the source code and start a fork
Please see SWOT.
While many Tiki developers run consultancy businesses around Tiki, yes, decisions are made in the best interest of the community, not of one or many businesses. For example, a business wants to find a way to lock you in. http://en.wikipedia.org/wiki/Lock-in Some could be tempted to make hosting more tricky so they can promote their own hosting offer. We want Tiki to be as easy as possible to install, via Fantastico panels for example. We are not just some geeks who are running Tiki on dedicated servers as root. We do care about the little guy on a budget shared hosting account.
Tiki is Community managed and open development with an official Tiki Software Community Association.
Tiki as a project, has modest expenses, which are covered by donations. It doesn't need a "business model" like a company that has expenses or employees to pay. While consultants are available for hire, Tiki is "Community Open Source" and not "Commercial open source". If you would like to hire or offer Tiki services, please see business models.
Absolutely! Various Tiki contributors can come and go. But, Tiki, as a project & community will continue no matter what. No one can lock down the code and require people to pay for it. Tiki evolution & development is in no way, shape or fashion the result of VC funding.
Users, just like you!
Typically,
Tiki is LGPL 2.1
Some systems have restrictions on removing promo link or logo. Some software is GPL but they ask money to remove the Copyright messages. With Tiki, you can do whatever you want. There is no commercial license. Since it's LGPL, you can include in commercial apps, sell, etc. Of course, we hope that you will want to contribute code, documentation, bug fixes, bug reports, feature requests, translations, or even money, etc. and/or hire some of the consultants, but you are not required to do so.
Yes, because Tiki is a community project.
In Tiki, that distinction doesn't exist. Or if you wish, all developers are core developers. We have a mods system to share code without being in the official release. However, this is the exception rather than the rule. Pretty much the opposite to many other systems...
This is a common concern/frustration of users. Especially when struggling with a bug.
Who would decide the priorities? Person A may think the priority is to improve feature X, but for person B, feature X is good enough, and adding Feature Y is more important. Of course, each person will tend to think their thing is the most important.
Tiki is a do-ocracy. Tiki is a diverse community. While many people have Tiki as their main livelihood, they don't work for Tiki. They work for various entities that use, love, (sometimes hate!), depend on, contribute to and want to see improve a project. But each has their own priorities. This diversity is part of the strength of the project. The answer is that different people that work on different things. The WishList has that name because it's things that people wish for. These can be bug fixes, feature requests, etc. The line between a bug fix and a feature request is often muddled.
Even if we somehow did stop people from adding new features (imagine the discussions!), would they really re-allocate their energy to fixing bugs? Since they may need a specific feature for a project, so if they are not allowed to contribute, perhaps they would use another CMS/Framework or they would do it anyway but not contribute. None of which is in Tiki's interest.
Now, to avoid regressions in the stable branch (3.1 -> 3.2), some developers have volunteered to form the Quality Team.
Why don't you draw the line and say, "this is what we do and that's it"?
Pretty much the same answer as "Why don't you fix/improve the existing features instead of adding new ones?" above.
Yes, but with care and consensus. We want to keep upgrades easy.
The goal is to have a useful application. Typically, features are removed because of the evolution of browsers or the web makes them obsolete. Or it could be because another feature in Tiki becomes so powerful that it makes this one useless.
In a traditional small-core-and-tons-of-extensions model, code will die off on its own by a lack of interest. In a wiki way, the community must proactively do housecleaning. This is much like wiki pages, which should be re-factored, deleted or merged. Please see: Endangered Features
Well, that depends on the people that participate in the community. So far, the number of features has been an asset, not a liability. Should this change, we'll adapt at that time. It doesn't matter how many features you offer, people still ask for more.
Of all the innovations, the rate of the totally new features has been slowing and what we see a lot is how various features are each improving and blending more tightly with other features. Ex.: Pretty Tracker.
You can shape the future of Tiki ๐
Predicting the future is very hazardous. However, as the past can often be an indicator of the future:
So, Tiki should be around as long as the technology is around. If you know what will replace the web browser, please tell us ๐
So Tiki will be around. But how good will it be? Well, the rate of evolution, innovation, etc. depends on the participation of people just like you ๐.
Thus, think long-term. Overall, we should be mindful about how our immediate actions fit in the long-term. Tiki will never be "finished" so any shortcuts now can come to bite us down the road (especially if someone else is stuck cleaning up the mess). See: Maintainability
Tiki has an all-in-one approach to features. Think of it as a suite of applications like Open Office. The features are built-in and all optional instead of offered via 3rd party modules/extensions/plugins. Because of this model (and the open participation), Tiki has more built-in features than any other Open Source Web app (if you know of one, please let us know!).
This can work only if you have the skills to tweak these applications to make them work together. And when each component is upgraded, you have to make sure the glueware is still working. And you need the programming skills to make this happen. If you don't complete the job, you will have a "Frankensite" patchwork of 3-4 apps, each with their own username/group & permission system. Each has its own template & menu system (if you care to unify the look & feel + navigation). Each will have its own search engine.
Tiki is a large application and there is a learning curve. However, it is easy to tweak. It was designed with simplicity & community development in mind. Please see the introductory documentation for developers and Coping with complexity.
That sounds like it could work and of course it can. However, this is open source software, not biology... So is it the most efficient? There are several challenges with this approach.
Competition is not the only path to innovation. Say you wanted to have the best possible definition for a word on Wikipedia and you have 9 volunteers to help. What would you do? Would you create 3 teams of 3 and pick the best definition? Or would you have all 9 people collaborate? If you split into 3 teams, chances are, when you evaluate what the three teams came up with, each team will have something you like. Yet, by picking 1 team and discarding the two others, you are wasting talent, motivation and energy. Now, it makes perfect sense to split the 9 people in smaller teams according to specific responsibilities towards a common goal. If this model makes perfect sense to generate content for Wikipedia, why is it not a good model for software development? Imagine 2 communities of the same size. One with the competition model and the other, with the collaboration model. Which one will have the most code review? which one will have more code/functionality duplication?
Feature interaction is a huge challenge with all software. In Tiki, it's not perfect but it's quite good.
Avoiding feature duplication doesn't prevent people from improving since participation is open. If there are several authors that want to improve a feature but have different needs/goals, they can just make them optional, and let end-users choose. It's free software. If you can code, just add the feature you need. If you can't code, try to find someone to help you.
Yes, but system Z has hundreds of modules. Extensions / 3rd party modules listings of popular PHP applications:
In the early days, search engines would boast on how many million web pages they indexed. And we had to navigate through pages and pages of results to find relevant stuff.
The important question is: "does the system do what you need now and for the foreseeable future". (This is just as true if you are picking Tiki vs another solution)
Tiki has over 1500 preferences/features/settings built-in, vs the sometimes 10 000+ external extensions/mods/plugins of some fine applications. But let's not forget that some of those external packages are features duplication.
The important thing is, do you see a feature that you want / need that is not offered by Tiki (please see feature list, modules list, mods list and all plugins)
One person's "killer app" feature may be useless bloatware for another.
Tiki is a shared resource. We want it to be useful for various people, to solve various needs, in various contexts. If it's not solving your problem, chances are, an enhancement would be useful to other people. The trick to try to use & extend the existing feature set. If there is nothing there that can help, it's to translate your need into something as generic as possible, so many problems can be solved with the same code, by changing labels or flipping a few switches. Of course, it's tricky to know this if you are new to the application, so you can work with the community to find solutions. A good example (of a generic feature) is trackers. Trackers are a database, forms and report builder, so you can create a bug trackers, a todo list, a mini-CRM, etc, whatever you want. You can build unlimited custom applications specifically tailored to your needs.
The real challenge is to find a technical and organizational way to deal with the Law of Software Envelopment.
Tiki8 has filters in the admin panels which will lower the learning curve.
Yes, it can be. If you don't use a feature, don't activate it! And http://profiles.tiki.org are there to help
Even if you think you just need a wiki or a forum now, isn't it nice that if you need a new feature later on, it will be there waiting for you? Just activate it, assign permissions and add links to your navigation.
This is called Dependency hell. The way Tiki is designed and organized, this has much less chance of happening. Yes, some Features can become unmaintained. However, they continue to work. If a developer enhances something in Tiki which breaks something else.
How many times in the 10+ year's history of Tiki have users ended up with data in dead-end abandoned code? This is very very rare (This could have happened with daily snapshots of SVN versions, but not when using stable versions)
While Tiki is self-sufficient, it offers many interactions with the outside World (services, technologies, standards, code re-use, data exchange, etc). Also, Tiki eats its own dog food.
Sometimes, it's less work and more future-proof to develop your own features. Just because Tiki doesn't integrate phpBB like many other projects out there, it doesn't mean we are reinventing the wheel. Tiki forums are very powerful and support threading, wiki syntax, attachments, etc.
In reality, Tiki relies heavily on External Libraries, to have less code to maintain and to leverage other projects experience. We try to contribute back (upstream) as much as possible. More than half of the code included in Tiki is from other FOSS projects.
Tiki 3.1 weighs 78 megs unzipped (24 megs zipped). Hard disk is really cheap these days. You can also use MultiTiki to have one set of files for several Tiki-powered sites.
Sure, if someone wants to do it ๐ Do you volunteer?
But it's really just simpler for everyone to just bundle all in one clean (albeit large) file. Many people install via Control Panels and thus, don't necessarily use, know or even have access to FTP.
In Tiki, we have a concept called WYSIWYCA. What You See is What You Can Access. The permissions system is very elaborate. Each user has a different experience, showing him/her what has changed since their last login (and only showing links for items which they are allowed to see). This takes more queries.
It would be interesting to have a real analysis to see if it's true. You would have to compare several apps with relatively similar content, features, etc. It's not that easy because apps are different. But let's say it's true -> Short answer for 98% of websites out there: who cares? Computers are getting faster & cheaper every month. People power is more expensive than machines. Yes, the code could and should be optimized in certain places. Yes, we should put some indexes on certain tables. However, we must always keep in mind that keeping the code simple makes it easier to understand, to customize and to fix. For the 2% of high-volume read-only sites: Please collaborate within the Tiki community to find ways to make Tiki faster. Ex.: http://varnish-cache.org/ Please also see performance for advice.
Bottom line: Systems that are heavily reliant on plug-ins are more apt to have security vulnerabilities for which there is no known patch. Why is this? Even frequently-used plug-ins are not always actively maintained and enhanced by their original developers. Source
Another risk: http://wpmu.org/why-you-should-never-search-for-free-wordpress-themes-in-google-or-anywhere-else/ Tiki (by default but can be changed by the admin) doesn't permit PHP code in templates, which is a nice benefit of Smarty.
On the other hand, Tiki has more code. And more code is more potential issues. Our solution to this is simple but effective: Each file in Tiki has a feature-check, which essentially renders the file inactive. If a security vulnerability is disclosed, admins can turn off that feature.
Both models have risks. We feel it's a good trade-off.
Every time people ask for an API, it's important to ask the question: "What are you trying to do?"
Tiki has no formal "API" but has several ways to connect/enhance/interact. It is designed to be easy to add functionality. Do you think over 200 people would have contributed via SVN if it was difficult? Hello World
Now, depending on what people are trying to do, they should look at Webservices, Plugins, Mods, Trackers, Interaction, External Authentication, etc.
Please read:
http://www.howtogeek.com/179213/why-browser-plug-ins-are-going-away-and-whats-replacing-them/
There is no question that the plugin model can and does work. This is to illustrate and explain how/why the all-in-one model works for Tiki.
During the development phase, developers can use Experimental branches.
Once a feature works (even if basic), it can be added to the main code base and tagged as Experimental
It is very risky to say something is better. The answer is it depends.
The Tiki model works for Tiki. And the community will adapt as needed. The Tiki model has benefits and drawbacks, therefore, the Tiki project has different challenges to tackle than other comparable projects.
In the small-core-and-thousands-of-extensions model, an important core competency is the Package management and the big challenge is dependency hell, which has prompted calls for Synchronized Releases.
Tiki has always, by its all-in-one model, had synchronized Releases. However, this brings another set of challenges. (There is no free lunch!).
Some things are good in Tiki and can be attributed to the model, but that doesn't mean other projects don't or can't have.
For example, Tiki has one full documentation for the whole application (over 1000 pages). It would be more difficult, but not impossible to make a similar resource with the small core/thousands of extensions model.
That is what some said about http://wikipedia.org/
Why would they possibly want to do that?
Answer: eyeballs
The quality of the documentation is variable. Some parts excellent, some are bad. There can be a correlation with the quality or activity level of the feature itself.
The good news is that it's a wiki, so please log in and fix it. If you know it's wrong but don't know what is right, ask in the Mailing Lists, forums or http://irc.tiki.org
This is certainly a possible model and likely used by some other projects. However, this is not the case in the Tiki community.
If you force developers to make unit tests, add documentation, or whatever, you will end up with fewer contributions. There is a valid debate to say "I prefer fewer contributions of higher quality". We are in the "commit early, commit often" mood and "things will improve with each version".
Some could argue that it's preferred to focus developers on adding/fixing features and it's up to the documentation team to handle documentation.
Whose job would it be to refuse (rollback!) a wanted feature because the developer didn't document? Can you imagine the debates there would be?
What often happens: Developers put up a skeleton documentation. Enough so that end-users get an idea of what it does and then, they can write proper documentation (hopefully with screenshots) from an end-users' perspective. Power users can also "interview" developers and capture this information to the documentation.
Tiki is the FLOSS Web Application with the most built-in features, is 10 years old and 250 people have contributed to the code base. It's not surprising that code quality and style varies.
Take any open source application and ask around if the code is good, is modular, scalable, etc. Someone will always find something to complain about. "critique" is very easy. I am sure if some of your work is available to public scrutiny, you will know what I mean ๐ The question should be: Does it work or not? Does it solve my needs for now and the foreseeable future? Can I improve it? Most people use Tiki without changing any code. So code structure/philosophy/organization is of no importance to them. They just want something that works and something which they can upgrade easily over time.
Also, Tiki was started with the technologies available in 2002: PHP 4, MySQL 3, Smarty 2, and no jQuery!
The good news is that various parts of Tiki undergo massive improvements. Ex.:
The all-in-one model actually makes it easier to do massive changes from one version to the next.
Please see: Modularity
Often, people with (or even not so) strong programming skills do not understand the unique challenges & dynamics of an Open Source project, especially a community-development one like Tiki. Some will design something which is supposed to be "superior", but it is too complex for people to use & participate to. Over 250 people have contributed to Tiki via CVS/SVN. That is a lot of eyeballs! The code is simple and has community coding in mind. Please see the goals of Luis Argerich, Tiki founder.
Please also see: I know this is all wrong, but fear it might be right
Tiki has been going on strong for 10+ years now. Tiki has more built-in features than any other open source web application (which is good or bad depending on your point of view).
Not only does it have a ton of features, it also has a very fast release cycle. Many powerful web apps have a release cycle of 1.5 or 2 years (for the core; modules evolves faster), whereas twice per year for Tiki.
There are many points of view over what is "professional" and "failure" vs "success". Some people could view that Wikipedia is not "professional" but it's hard to argue that it's a failure ๐
For many years, some people have expressed these concerns. There was a lot of FUD being spread around. Perhaps Tiki is like Wikipedia and is "is one of those things impossible in theory, but possible in practice.". However, here we are. Tiki is 10 years old. It's a mature, very powerful application, with a large community of contributors, with tons of features and which is tons of websites & Intranets. Of course, there are a lot of things to improve and there will always be. Remember: if you don't need a feature, don't activate it. However, if you need it in the future, it's only a few clicks away. And remember also, Tiki is open source and open development. YOU can participate in its future.
And Tiki just won a Bossie!
http://www.infoworld.com/d/open-source/bossie-awards-2010-the-best-open-source-applications-150¤t=9&last=10
1) |
19 Sep 2024 14:00 GMT-0000
Tiki Roundtable Meeting |
2) |
Tiki birthday |
3) |
17 Oct 2024 14:00 GMT-0000
Tiki Roundtable Meeting |
4) |
21 Nov 2024 14:00 GMT-0000
Tiki Roundtable Meeting |
5) |
19 Dec 2024 14:00 GMT-0000
Tiki Roundtable Meeting |