OK, I think we have, yeah, the keynote is over, so everybody has to have time. Welcome, my name is Paul Roulant, I am a sysadmin and an integrator, and this is going to be a rant, more or less, but with a positive outcome.

Fifty shades of Plone, because Plone hurts us in many, many ways, and now I am not even talking about PloneCore, I am talking about Plone add-ons.

They can be a bit of a pain, and although some people enjoy pain to each their own, but there must be ways to do this better, and this is going to be a call to arms to do just that. First of all, I am going to try to make an analogy, and I know they always suck.

But then again, analogies can make things a bit more clearer, and whether this will be a good one or not, we will see. So, we are going to go with bad analogies, and I am going to use the theatre

or the movies as my analogy, and that add-on products are like actors in a play. Think about that for a minute. You have a play, you need actors to actually put it on. So, what does that make the integrator? You get a cookie, sir.

So, integrators should think of themselves as directors. You are directing a bunch of different things, and hope the outcome won't be a big disaster. Still, that means work, because whether you are doing a serious thing or just a musical, one actor is never enough.

You might have a monologue, but they are a sub-genre. Most of all, mostly you need more things working together, so you need an ensemble, you need products working together. But yeah, being a director sounds really great. You get all the dazzling power of a Hollywood mogul.

You get like, ooh, there are so many actors to choose from, and actually you are allowed to typecast. You can say, I want a bad guy, and he must look really mean. I want like the heroine, and she must be beautiful, but yet tough. So, you have this immense power.

And if you just start out, there are tons of new actors coming in who don't have money, who want to become famous. So, it's great you get so many to choose from. So, you are in heaven? Maybe not.

Because, yeah, as soon as you audition for the bad guy, you get like 20 you have to choose from. You have some kinds of add-ons or of actors that take on a life of their own. Suddenly, they turn out to be musical artists where you didn't want that.

Some of them do weird stuff that you really don't want them to do. So, it's all starting to look a bit less nice. And at the end, you are going completely insane, because once you have them all in the same room, they throw hissy fits at each other.

They refuse to work together. They are like, I can't work with this person anymore. They want weird stuff as requests. They want weird versions of software. They want cocaine and drinks and stuff in their dressing room. It's horror. And, worst of all, suddenly, they start interpreting what you want.

You say to them, give me like a scene of true romance. And suddenly, they think they know better and they start like, yeah, it must be edgy. And you are like, no, it shouldn't be edgy, it should be romantic. So, they take on a life of their own. Products often do that. And also, yeah, products are indeed like actors.

They come in many different sizes. You have like the young and the stupid, who may look pretty on the outside, but actually, once you ask them to do something real, it becomes like a nightmare. And then, you have the old and the bitchy ones, who know exactly what they want,

but you really have to live according to their rules. By the way, this is Joan Crawford. If you ever come across a copy of her famous book, My Way of Life, it's been out of print for decades now. But occasionally, it turns up, get it, it's the most insane thing you will ever read.

She fills it with handy household tips for dinner parties like, don't forget to throw in fresh water lilies in your swimming pool. Who would have thought of that? And like, always make sure that the personnel is sort of cleaning up the dishes at least seven rooms away from the dining table. Otherwise, you hear them washing up and that's annoying.

So, practical everyday advice we could all use. So, Joan Crawford was not only famous for being quite annoying, she also had a daughter who didn't think much of her, who grew up in a Hollywood household.

And growing up in Hollywood households is like products being thrown out into the public too soon. Child stars, they sparkle for a moment and then they go really, really batshit crazy. It happens. It happens to a lot of products.

It happens to the product ecosystem and you end up with a wrecking ball and a mess. This happens if you have like a Disney or if you have a Disney child star. It happens to the best of us.

Okay, next slide is, I have been accused by several people actually present in this room of forcing people to use cat pictures in their slides. So, I'm going to give you one kitten. It's a moving gif made from a vine of a kitten doing a Miley Cyrus interpretation.

That is the internet in one slide. No need to ever turn on the internet again. So, but enough of the fun. We are having a problem with our products. We are having a serious problem and we need to solve it.

So, what has been the way rational people think, what has been the way to solve a problem? We turn to science. Science, cold heart science. We can't just go randomly picking solutions. No, we need to think about this.

So, what are we going to do? We have a problem. We have more than 1100 packages now on the collective dot star name space. That's not all. We have a lot of other products. It's like actors again. We have the Hollywood ones. We have the Bollywood ones. We have the Nollywood ones.

We have Chinese stars. We have Korean soap stars. There's just way too much. We can't get it right and some of them will never see the light of day. They'll go into the straight to video markets and that's a waste. So, and the problem is new ones, new products come out almost on a daily basis or refuse to die.

So, the problem gets worse. And as you can see, it's an exponential problem. The more products you have, the probability of you getting the right one rapidly declines to zero. Actual data measured over a long time.

Science. So, this is, yeah, this is not sustainable in the long run. And it gets even worse. We have things like coat rods. We have add-ons that haven't been maintained and that you can't use in the newest, shiniest clone version, which is quite annoying.

We also have what I call the life of Brian syndrome, where you have the people's front of Judea versus Judean people's front or collective carousel versus products carousel or collective doormat versus products doormat. As an integrator, you're like, what?

Which one should I pick of these? And we have what is also a problem, the jQuery Jenga, as I would like to call it. jQuery is like this big pile of blocks and as soon as you touch it, it all falls down in very interesting ways, never repeatable, because everybody wants their own little version of jQuery

and it's just a nightmare if you're running aside for a longer period of time. And that's not even the end of our problems. One of it is also if you actually try to find a product and you start looking at documentation.

This actually is actual documentation taken from Plone.org product center, where somebody... I made it small so that you can't read it, because the author would probably be offended,

but there were a lot of them to choose from. The only documentation provided on Plone.org were doc tests. Now, doc tests were maybe a good idea when they were invented, but let me tell you, they're neither documentation nor are they tests. I don't know who came up with that name, but if you're still using them, stop.

If you must use them because you're addicted to them, don't think this is your user documentation, because it's not. Especially, this was part of that same documentation, it also includes the errors.

So this is what... It's a test, all right. So people are like, okay, this is a product, what does it do? Oh, it breaks. Hmm, let me get that one. Great advertising, folks. This really helps.

So any integrator ever will have this as the reaction to seeing that as documentation. It's just like, seriously? This is not helping. And yeah, the end result is just a bit sad.

Somebody put in a lot of effort to write probably a great product which does stuff. We might never know what it does because nobody ever bothers to try it out. And you end up with zero likes and zero dislikes. That's a lot of effort gone to waste. And that's a bit sad. I'm going to pull my sad face now.

So, stop. Really, people, we have to stop doing this. We must do better. And in order for that, we go to one of the soft sciences. They're sort of a science, psychology. We all need that.

And here in this helpful diagram, you see the stereotypical, it's a bit over the top, it's a stereotype. But we see the split between developers and integrators. Developers want to do cool stuff, solve their own problems, be seen as clever people who do stuff.

And then they say, please leave me alone with your customer request. And move on to the next project, the next interesting problem, ASAP. That's in the nature of developers. I don't blame anybody, but that's sort of how it goes. Integrators, on the other hand, they want to have all that cool stuff because they have sites to run.

They have users in the end turn, like real people who come at their desk and shake them. And say like, this doesn't work. They want to tweak and fiddle. They want to upgrade over the years. And they want to adapt the products to do exactly what their end users wanted to do.

So, and they're almost entirely separated in their mindsets. The only thing, as you can see in this scientific diagram, the only thing that unites them is the love of Plone. So, for the love of Plone, please people, we've got to make this better.

So, we're going to raise the standards. We're going to make it as easy and pleasant as possible for both devs and integrators. It's important to keep both sides happy here. And we're not going to do it by mass producing as many add-ons as we can. Because, well, you could. You're free to do so, but it's not happening.

It's like churning out yet another beer variety of the one that all tastes like Heilinken anyway. There is no point. We want high quality. We want things that people really take care of.

Things that you really look after and that you're happy to use. Those are the products that we should really cherish and that people will love. So, how do we get from the current pain to the pleasure of carefully selected, nice products?

We need a plan. To get out of the current, I don't know how many thousand products to a better solution, we first need to pick and praise the best ones. We need to say, hey, these are examples, shining examples of things that really make the life easier

of any plan integrator or anybody who runs a site that actually solves problems and actually also documents how it works. So, in the pick and praise process that we're going to do, we also have to then, of course, explain why we pick certain products,

like why these are better than others. It's always been a contentious issue, but I think we're reaching the point where we just have to say some are better than others. And, yeah, we might as well have some fun in the process because if it's not fun, it's hardly worth doing anyway, is my motto in life.

But remember always with? Science. Yes, you've got it. Whew! So, science applied to picking one thing out of many is actually quite easy. It has been done before. Stick any drag queen in a thrift store and he or she will get the best stuff out immediately.

It's a scientific method of scanning the room laser-like and zooming in on the right ones. If you know what you want, you're going to get it. So, if we're going to science, we might as well go to applied science.

And the biggest science laboratory in the world is TV Land. TV Land has billions of viewers. They have massive audiences. They have a profit motive. TV producers and TV selectors are never wrong because they measure shit.

They do it with science. So, in the ratings as a handy bubble chart, you can see anything else, news, documentaries, series. They all get sometimes good ratings, sometimes they don't. What consistently wins is the talent hunt in all its many varieties.

So, scientifically proven, we need to have a talent hunt. Then there is the question, you all know talent hunts. It's like country X has talent. The voice, dancing on ice, dancing on polar bears.

I don't know, they come up with new stuff every time. So, then we need to decide how many judges do we want. Again, science. It's been proven in many countries that the amount of judges correlates directly to the popularity of the TV show.

And as you can see, it's clearly obvious that we're going to need four judges. And we're not doing this just for fun. We do have goals with it. My goals for getting this talent hunt on the road,

which will be quite soon, is that new people can start sane. They can actually look at the products and say like, oh yeah, I want an add-on that does this. And they pick that add-on and they can be pretty sure that it won't blow up their site instantly. It also has the advantage, once we have picked a number of good products,

that other authors can look at it and say like, ooh, so that's how you get into the top 50. Hmm, I might learn from that. That would be ideal. And most of all, what is my, also one of my goals is that we improve Plone's marketability. Because in the marketplace, good add-ons count.

Because in the end, you sell to stakeholders, managers, team leaders, not extremely techy people. It's kind of rare that the top tech person in a client chooses. In the end, it's always some kind of manager who decides what system they're going to use.

And you know what? All managers nowadays now know how a good, well-tended app store looks like. They're going to look at Plone.org. They're going to browse the products, and they're going to expect sort of like an Apple or an Android-like experience, where they can easily flick through things like, hmm, we could add that.

We could add that. We probably wouldn't even need an integrator to do that. They're going to look at it. If they're going to be presented with like 2,000 random products with crappy documentation, that's going to count that. If you have like a nice store, that actually helps improve our marketability and our selling point a lot.

Because happy managers mean more Plone. And also, I think documentation should be taken serious, like really serious. We don't actually, or any sane person nowadays in our community

doesn't allow new code that isn't tested. We have tests for just about everything. We've done unit tests. We've done integration tests, robot tests. I think it's time for the next level, and that is test on humans. Peter will be fine with it. We're not testing on kittens. We're testing on humans.

So that should work. And these are the two things that everybody should get into their mind. Undocumented code is broken code, just as much as is completely untested code on the code testing level. And if your code can't be found, you probably have a hobby,

but nobody else is really profiting from it. It's not very useful. So in this race, integrators, in this talent, integrators will play a key role because as add-ons go, integrators pick them and they use them. So I'll tell you some secrets about integrators that will help you write better products.

First of all, we uninstall. We do quite a lot. So it's really great if a product actually can be uninstalled and still leave your site running. I like that. That's good. We use more than one product, amazingly enough.

So if your product sort of takes over an entire site, that's kind of bad. It should be as non-invasive as possible so that other products can do other stuff. We like turning dials. And we're also fine with line fields.

By that I mean don't invest your time as a product author to write a super sleek control panel. Integrators are really fine with doing stuff in the registry. We don't mind. We're not end users. We can do that. Invest your time in documentation and not in a super shiny control panel.

If you want to have both, fine. But if you have to choose, if you're time limited, integrators can live with a few warts and bumps on the user interface. We're fine. We also have some darker secrets. We suck at CSS, but not as bad as you.

So all we want from product authors to do is provide some basic CSS and classes to hook our Diazo themes into. That's all. Your product doesn't have to look great because we will theme it anyway. Just provide hooks will be fine.

We won't like your jargon choice. That's because we also don't like our manager's jargon choice, but she pays. So if you can, try to put as much of your vocabulary in your products into translatable files, into PO files that we can fiddle with. Because if you write your add-on and you use all commercial jargon,

I'm pretty sure my nonprofit will say, we don't have clients, we have partners. And the other way around as well. So people will have to adjust the language anyway. So make sure if it's somewhere in an easy to find place so that people can relatively easily change it.

There has been some thought on how to make translations changeable through the web, which would be an enormous win for just about everybody. I think that's a really good idea that we all should think about. We'll use your product in weird ways, which is good. And we upgrade.

So make your upgrade steps repeatable without harm. If your upgrade step blows away my catalog when I do it twice, I don't like that. Because we upgrade a lot and sometimes you have to do it in order to avoid the jQuery Django.

But it's time to get back to fierce competition mode. We're going to have a talent hunt. We're going to do it by jury. You can also nominate, but it's going to be a jury only way. And no televoting. If you have too much money and you want to spend it on stupid stuff,

donate to the Plough Foundation. I'll tell you a few of the requirements or the conditions that we're going to look for in the products. These ones are non-negotiable. If a product doesn't fulfil these, you're out.

Next. It should be on Plough and PyPy, have a repository, be translatable, have an actual readme. And if you can, please, in the first three lines of that readme, tell me what your product is supposed to do. That's really handy.

And you'll be amazed. I've looked through almost all of the products. You'd be amazed at how many just seem to assume that you already know what this product does. People don't. They're looking for solutions. They're not looking for your specific name. People are not pre-scientific.

You need to tell them. You get bonus points for things like screenshots if you link to a real-life site that uses it or even if you blog about how you implemented it. You also get bonus points if you compare with similar add-ons and say where yours is better or different

and so it's a different use case. By showing that to me as an integrator, it shows that these guys have actually thought about things before they wrote this product because they are aware that there are other ones doing similar stuff and they actually point to that so they must be pretty confident of their own solution.

That's a good sign for me. And an issue tracker where things actually get answered is always such a vote of confidence. You can, of course, always get more charities on top if it works with both dexterity and archetypes, if there are already translations, if it's properly tested.

If you actually tell me what are the potential problems with your product, ooh, big bonus. If you say, like, use this for a site up to this and this, many users after that, hmm, I'm not quite sure. Thank you for telling me in advance. That's really good.

You'll be dipped in triple and thrown to the Plone Conference if you include any of these, if your product is Plone 5-ready, if it has a screencast, and if it has many configurations. So we still have room for two more judges. There are already two judges. You can come to me somewhere during the conference.

We want people that can do different stuff. It's not that difficult. And to win at this, developers, you are allowed and your occurrence to cheat. Read the rules, improve your documentation. You will have until December the 1st to upgrade your products with great documentation.

So do that. Integrators. If you think, like, ooh, I really love this product, but, yeah, I know the documentation sucks, what is stopping you from actually helping out the developer? You've had so much use from that product. It's time you gave something back.

Help document it. Write a blog post saying I'm using this great product. Here's a link to my site. You would help them in the competition, but you would help them in general life, and you would help Plone in general. So you can be on team collective cover, or you can be on team uploadify, or choose your favorite, and you can play for the team

and make sure that add-on gets in there. And this all ties handily in with the big Plonark refit, because the Plone Software Center will be gone. Hallelujah! If you're still using the tracker, get out.

The new one will have a much simpler Plone product showcase. My dream is that it will be tagged, and you can do faceted searches on it, because lots of products fill multiple categories. The idea is that you write once about your product,

and you don't have to update it once a new version comes out. It pulls that from PyPy. It's just like a blurb saying, I have this great product, here's what it does. Don't bother to put your integration products on there, just leave them on PyPy. If you have a bundle, if you do something like Plone Social, which is composed of many different things,

write one, say, I have a bundle, and all the implementation details, leave them on PyPy. We're still thinking about if this should be curated or not, but that's a discussion that should be held by a wider section of the community. In terms of marketability, it would be really good if all the things on there were actually readable

and present us in a good light. Integrators can help. They can get credit for doing it. Product authors can allow other integrators to write their documentation. We have a content management system, you can do that. You can delegate responsibility.

And of course there will be people that say, that's a really bad idea, because they really need their gallery. To that I say, be my guest. Just make it better than the previous one, or at least different enough. If you say, that's against the spirit of open source to sort of not allow just anything on your site,

I'm like, stop whining. Nobody is going to stop anybody from putting products on Pi Pi, on the collective, on GitHub. It's just like we want something presentable on Plone.org. And if you can't write docs, well, get in touch with your users.

They can really help you. And some of them might actually be glad to give something back. So we're on to world domination. We'll have a better product showcasing. It'll be in line with our marketing messages, which are going to be revealed pretty soon, I hope.

Yes. And we're going to kick start this process with electing the fabulous 50 products. I am, as I said, still looking for judges. The idea is to have one month of judging starting December 1st until December 31st, and then in January announce them.

The top 50 products in different categories. We will crown them, shower them with roses. We will not call them the Plonys. If you mention the Plonys, you will be killed. We will find another name. And hopefully by that time, we'll have Plone.org ready also to showcase them with a little seal saying,

best of 2013, or something of that, because we might repeat this next year. So, ladies and gentlemen, the race is on. And let's improve the products. Thank you.