[Product-Developers] Re: How can we win the battle for developers?

Dylan Jay gmane at dylanjay.com
Fri May 9 01:34:04 UTC 2008


Martin Aspeli wrote:
> Dylan Jay wrote:
> 
>>> I think there are two ways to look at this:
>>>
>>>  1. You can't find good developers because no-one wants to do Plone 
>>> work.
>>>
>>>  2. You can't find good developers because Plone is growing and 
>>> everyone is
>>> busy.
>>>
>>> If you look at plone.net, there are over 250 companies that offer Plone
>>> services. I don't think that's a sign of weakness.
>>
>> Not weakness now but potentially in the future. I think there a lots 
>> of companies using plone because they have settled on the best 
>> technology available, ie Plone. It delivers. We all know that.
> 
> I wasn't talking about companies using Plone (there are many more of 
> those); I was talking about companies offering Plone services.

So was I.

>> But my argument is that if all the fresh blood is doing into 
>> ruby/django/drupal then they will be reinventing the wheel there 
>> instead of discovering Plone and improving it.
> 
> Sure, that could happen, and it would be unfortunate.
> 
>>> Plone is never going to beat PHP for attracting college kids who want 
>>> to put
>>> together a quick dynamic website (not that this is all PHP is good 
>>> for, of
>>> course). That's not necessarily a bad thing, though. Most people who buy
>>> Plone services also want to deal with professionals with some 
>>> experience. I
>>> think Plone has a place in the market for skills and investment, and 
>>> it does
>>> reasonably well there.
>>
>> Professionals all started as collage kids. And many professionals are 
>> hobbiests who want to play with technology in their spare time at 
>> first, but that soon terns into a serious skill and that turns into 
>> contributing to the community or offering their services to others.
>> These are the people we want to attract. These are the people who will 
>> use google app engine and django to create some play apps. Or use 
>> drupal and start telling everyone who extensible and powerful it is. 
>> and then convince their corporate IT to use drupal for their intranet.
>> For these reasons I think we can't be too arrogant about what kind of 
>> developers we attract and how many.
> 
> I certainly don't think we should be arrogant. But I'm not sure you've 
> captured the whole market with this analogy. People who make technology 
> decisions usually make them based on more than what they happen to have 
> tinkered with. They look at things like track record, scalability, 
> availability of support and feature sets. By your argument, no-one would 
> be building applications on J2EE. ;-)

They make those decisions on either coolness or availability of 
resources IMO. Thats a big simplification I know. J2EE is popular 
because you can get java developers and you can get java developers 
because its a good career choice for a developer to learn. It's a 
chicken and egg thing. But where do new technologies come from? They get 
created and then get a buzz because they are cool. They are easy to 
learn and very functional, productive or whatever. (or microsoft creates 
them and so you know you can get a job if you learn it)
Imagine grok becoming the new rails and by learning grok you learn most 
of what you need to customise plone sites. Thats a nice aim.

>> I noticed in the summit summary some words about hosting. I think this 
>> is pretty important for attracting developers and we can do more in 
>> that area. I don't think any action items got created for this.
> 
> See http://dev.plone.org/plone/ticket/7837

static html deployment is one option. but it cuts all the dynamic nature 
of plone away. It would be fine for people who just want a simple site 
but not of interest to developers who want to get dirty with customising.

>> I was thinking about how easy we could make plone deployment.
>> What is it people want to know? How easy is it going to be and how 
>> cheap is it. With VPS prices coming down you can host a few plone 
>> sites for $20 a month.
> 
> I don't think that's viable, unfortunately. You're looking at more like 
> $100 a month for something that can serve a moderate site. That's still 
> cheap, of course.

Not true. Rails has created this ecosystem of xen based VPS providers. I 
use slicehost.com for instance but theres highspeedrails.com and plenty 
of others. For $20 I get 256RAM and ubuntu with full root access. Thats 
plenty.

>> You just have to
>> - learn some unix admin.
>> - and find a VPS hoster.
>> - work out how much memory you need
>> - and setup a buildout on your machine
>> - and put your code in svn
> 
> You don't need this, of course.
> 
>> - and deploy it to the server
>> - and then buildout on the server and then run.
>> Buildout is awsesome and that is so much less than it used to be. but 
>> can we do more?
> 
> See the link above.
> 
>> Can we make google app engine experience that will work on an 
>> ecosystem of VPS hosters?
> 
> I kind of doubt it. But then, I'm not sure I agree that this is such a 
> hugely important factory. Deployment could be simplified, for sure, but 
> I'm not sure that's where it hurts the most, by any stretch.

learning curve is the biggest problem I think we all aggree. but part of 
the learning curve is getting a functional site on the web as well as 
customizing. Maybe all thats needed is a really good tutorial. Perhaps 
thats a good weekend task for me. "how to get a live plone site in 10min"

>> We need a server builder.
> 
> Do we? :)

glad it got someone else thinking about it :)

>> Download an installer. It runs the buildout builder and then the 
>> server builder. Pick from a list of supported Ihosting companies. Hit 
>> "go" and you have a complete development environment and deployment 
>> environment all setup for you.
>>
>> Just an idea.
> 
> I think that'd be cool, but realistically, that's going to take a *lot* 
> of setup and relationships with hosting providers. I also think that for 
> serious sites, people are going to want something more bespoke over 
> which they have more control.

actually we don't need to list of hosters. All we really need is some 
very simple instructions on what to do with a brand new ubuntu install 
to get plone up live. We almost have that with zopeskel and buildout. I 
just wonder if we can do more to connect your development environment to 
your hosting enviroment.

I'll write my tutorial this weekend and we'll see what we can simplify :)

> Martin
> 





More information about the Product-Developers mailing list