 All right, I think we can get started Hi, everyone. My name is Moshe Weitzman. I'm a long-time Drupal developer and consultant and member of Acquia I want to talk a little bit about our web hosting requirements and Perhaps make a larger conversation about what is Drupal's target audience? Hopefully everyone's not too fatigued by now about that question I'm feeling a little bit fatigued about it, but This is a group conversation. This is the only slide in the presentation We can do this and And you know everyone can talk So I think most people can agree that Drupal has an extremely wide target audience From a hosting perspective We definitely cater to people who are on $5 web hosting, you know the cheapest PHP Web hosting that's out there Where you hire, you know very highly paid consultants to set up multiple levels of caching and servers and CDNs and multiple of everything and It's really quite an amazing range there And so that's great Except for all of the negative effects What one of the things that happens in that case is that the core development team has a really Hard time managing all of these different audiences And so you end up having burnout in the in the Core development team. We've also heard a lot about burnout in the last couple weeks in here I can think of an issue in Drupal 7 which really Burned out like a top developer Barry Jaspin came in worked hard on it And went away Maybe he'll come back, but but he's gone and that's a huge loss for a core development team The one I'm thinking of is in Drupal 7 you are not allowed to delete text formats You're not allowed to delete text formats that have any content that's affiliated with them and the reason being that it's really hard to find all that content and Delete it in the context of a web request even with our batch API and So the solution to this long painful issue was well people just can't delete the stuff and We don't even have a way to tell you how to go about deleting all this content So that you can delete your text format. So if you create it, it's there forever In other All right, so what would it look like to bump our web hosting requirements There are three things that I put in the session description Some people haven't seen it. So I'll just repeat them Let's say that shell axis is required in order to effectively run Drupal So you need it to be able to SSH to your site Well that lets us take advantage of dress scripts to manage your site So deleting a text format now becomes fairly trivial thing because we're not subject to timeouts We can go ahead and do it and you know, there are lots of Other things this would solve There are some other features that we might give up if we know that people have shell axis like the Install a module through the web thing which you know sounds great Unfortunately the way that it's implemented it requires a lot of your server namely you have to have the FTP PHP extension Or this SSH PHP extension which hardly anyone has so Effectively that feature is neutered by its security requirements So, you know one other point I want to make about The web hosting is that our current range was really developed in a time before Where we are now, which is that VPS is and virtualization is super prevalent You can buy a VPS for Ten dollars, I know you can definitely get it per month. Maybe they have it for five But the point is that's not too expensive anymore Naturally, it requires a bit more This had been experience to manage app to get is not hard to use but The point is it's more than nothing And The other web hosting requirement is that level would be The Development teams can assume that you have a PC installed when we are writing software which Impacts exactly, you know the things we care about And so We had you know huge discussions about this function registry during the development of triple seven and One of the benefits not the only one was that it was really great for people who did not have a PC and it was Less important for people who had a PC and so a lot of development cycles went to something that a lot of Drupal customers Wouldn't benefit as much from that eventually got rolled back. So it's a little bit of a moot point now And you know, I think most people are happy it got rolled back But that's something to think about we raced in a lot of cycles there if we could assume that everyone had a PC Then we may not have proceeded with that project All right, well, I think that's enough for me. I want to hear from other people about target audience and web hosting requirements and We'll see where we go Jeff I Guess one other thing I want to introduce the conversation then I'll get to the hands that are up I'll remember where you guys are you can take it down is that at Some point on the low end These people might want to look at Drupal service providers. Okay, so Gardens is a big one buzzer is another one You know people like hot Drupal who specialize in Drupal hosting Our services that weren't around a few years ago when we made this broad range So I think if you get low enough, there are people who are very happy to take your business And they actually do it with varnish and memcache and stuff that you would never set up for yourself So you might actually get a Better experience and still pay not a lot of money All right, so let's go here That's right. So VPS is really a poor term for what I mean. It means that whoever's hosting your website Can deal with Drupal's increased requirements web hosting requirements. Yeah Okay, so that's and not running locally means that there's not a point of presence near you or yeah, okay You know, you know one approach is well the market will provide that right? I mean we have lots of entrepreneurs here from Australia. Maybe they will make something that appeals to that audience, right? Okay, that's And good point right here The suggestion is to More quickly upgrade our php our base php requirement. It's 5.2 and Drupal 7 I guess the suggestion is to go to 5.3 for Drupal 8 It's already 5.3. So I don't think we can go higher than that since 5.4 is an alpha Yeah, they usually make both available right for some period of time Okay You had a chance has hurt anyone else Right in the back Yeah, I mean, I think it's a good distinction It's a good distinction to say that your production server might be the one that has to have increased requirements, but not your devs It is amazing, but it's true they don't have it and Every performance optimization you might think of it pales in comparison to that one. It's more for the developer The developer understanding and focus when we broaden the talk to what is Drupal's target audience and Should it have a narrower target audience? Would we thrive that way? You know, there was some interesting data about We're losing the raw number of sites, but we're gaining the high-end sites Would be we'd be happier with 50% of the sites we have now, but it's more focused group like mid-range and high-end clients, you know, so It's true. I mean, it's something to consider and How you got here? Is it necessarily where you should always stay? But you know, that's not something you turn your back on lightly Yeah, all right other hands right here Right, so be mindful of all the nonprofits to use it all the way in the back Okay, I don't think that at least the requirements I mentioned are in that category Like you don't have to set up memcache in order to run Drupal. That wasn't one of the proposals But somebody could suggest a requirement that inhibits the development environments here Jakes we talked about that a little while ago I don't know maybe you weren't in but we did mention that there are service providers And there will be more that could handle that class of customer Okay, right back here. Okay, we we don't like it when people copy paste commands Yeah, and APC even if we got that it would just be a recommendation because it it really doesn't do anything If you did if you said worked on a web server that didn't have APC all this everything would work It would just be slow Right, so anything that could reasonably be done in a browser Drupal ought to have it be available on a browser, right? I think everyone would agree with that but something like deleting a text format I think even you suggested like those sorts of things. Maybe we could start relying on the shell to do them Right right, I think the problem comes in when We spend lots and lots of cycles doing something that's much better done on the shell and that You know download a new module and install it thing in Drupal 7 would be an example of that Yeah, I mean seems like something that we ought to do in any case. Yeah, I Thought of another example that I'll share During the Drupal 7 release cycle we had a push to put poor man's cron into Drupal core. It's on by default And It does something that's pretty crazy. It like makes a new request back to Drupal via hidden image and it it goes ahead and calls Something that cron dot php also does Or I guess it calls into cron dot php That went into core eventually and It had been implemented in the worst possible way it you know the second every single request in the Drupal Caused a second request into Drupal That did a full bootstrap to check if cron was needed to run And if it didn't it would stop there and if it did it would run cron so a Lot of core developers had to go now fix this thing that had toasted performance in Drupal and You know that was another saga so When we start doing stuff that's better done in the shell We really burn out our core development team I remember the patch was a good patch, but didn't go others Right here Right, maybe distributions are part of the answer. Maybe some distros can have different requirements and others Do you have a hand up Narayan? No, okay, right here. Okay. Good point right here There's a notion that your website could be in a mode like this is a production server And maybe we would give messages based on that here Yeah, I mean, I think there's there's truth to that if you have the answer for how core development should come to agreements that would be cool All right good point Jeff Can you elaborate? All right, I think that we've out of we run out of time for this core conversation. There's another one right behind me So thanks everyone