Acceptance of ZF2

classic Classic list List threaded Threaded
28 messages Options
12
Reply | Threaded
Open this post in threaded view
|

Re: Acceptance of ZF2

Tim Fountain-2
On 10 June 2013 15:30, Matthew Weier O'Phinney <[hidden email]> wrote:
On Sun, Jun 9, 2013 at 6:08 AM, Tim Fountain <[hidden email]> wrote:
> If you search for anything ZF2 related on Google, results that point at the
> official site mostly point at version 2.0 of the documentation (or sometimes
> 1.12). This is bad, and is going to become more of a problem over time. By
> always having the current documentation at the same URL, that ends up being
> the one people will link to from blogs and the like, which means that'll be
> the version most like to appear in the search listings.

This one became impossible when we added v2 to the mix. The reason is
that any pages that were hitting "latest" would have been instantly
out-of-date, and the search results would have been incorrect.
Essentially, when we have multiple major versions, each getting
updated versions periodically, what is really "latest"?

Yes, I can see the problem. I guess the question is, is it better for search engines to always link to the current version but external sites to potentially end up with mismatched/broken links, or for external sites and search engines to both potentially link to older releases. 

I think we could have /1/ and /2/ links, however, which point to the
latest minor version of those respective major versions. Would that
address the issue, Tim?

I think that would help a little, but if the plan is still to do major releases a bit more frequently then we'll still have the same issue in the long run, just with major releases instead of minor ones.

I've just been looking into how other frameworks handle this, particularly those that have been around for a few years. Most seem to have version numbers in the URL, and many also have a 'latest' or 'dev' one, but they vary as to which one they link to primarily.

The solution I like best is Django's. Their docs work in the same way as ours (version number in the URL, website links to the latest released version). But they have an XML sitemap which uses the 'priority' field to influence which one the search engines link to in the listings: high priority for the 'current' version, low priority for the older ones. I think this is probably the closest you can get to the best of both worlds. I've just checked out the source for the zf site, I'll see how easy/difficult it would be to implement this.

--
Tim Fountain
http://tfountain.co.uk/
Reply | Threaded
Open this post in threaded view
|

Re: Acceptance of ZF2

Ralf Eggert
In reply to this post by weierophinney
Hi Matthew,

> First off, from download statistics, we've seen a huge upswing in ZF2
> adoption -- to the point that we now have a 60/40 split ZF2/ZF1
> download rate. Yes, you read that correctly -- ZF2 is being downloaded
> more than ZF1 now, and actually has been for around six months now.

This sounds great. Do you have any absolute numbers you can publish
officially?

> [...] As several others have
> replied in this thread as well, if they have a working ZF1 app,
> there's often no compelling reason to upgrade to ZF2 -- why dump
> effort into an upgrade when it delivers what it needs to perfectly
> well? [...]

Yep. That is the main reason why most of my customers stick to ZF1 since
they don't see any economical reasons to switch.

> One definite issue, however, is how to find developers with ZF2
> experience. [...]

This is a aspect I did not thought about yet. There must be some good
reasons for it. One reason might be that there is no English ZF2 books
out yet. There is one German ZF2 book out (mine ;-) but I really wonder
why the English market for ZF2 books is not existing yet. So, English
authors, where are you?

> Documentation is often cited, and yet I hear about an equal number of
> folks saying the documentation is horrible as are saying it is
> wonderful; I think it's next to impossible to have documentation for a
> project this size that will be universally useful. That said, we could
> and should have better documentation -- the problem is identifying
> what needs to be done, as well as getting enough attention diverted
> from the fun tasks of coding to the tasks of writing docs -- which are
> very, very different skills.

Yep, I always hear people bashing the docs, but when I ask them, what
should be done, I get no real suggestions. Ok, some components are
missing IIRC and some component docs could be improved. But maybe people
are looking for more docs on how to combine all components to a full
featured ZF2 application. The user guide is only the first step.

Regards,

Ralf
Reply | Threaded
Open this post in threaded view
|

Re: Acceptance of ZF2

David Muir-2
On 11/06/13 06:55, Ralf Eggert wrote:

> Hi Matthew,
>
>> First off, from download statistics, we've seen a huge upswing in ZF2
>> adoption -- to the point that we now have a 60/40 split ZF2/ZF1
>> download rate. Yes, you read that correctly -- ZF2 is being downloaded
>> more than ZF1 now, and actually has been for around six months now.
> This sounds great. Do you have any absolute numbers you can publish
> officially?
>
>> [...] As several others have
>> replied in this thread as well, if they have a working ZF1 app,
>> there's often no compelling reason to upgrade to ZF2 -- why dump
>> effort into an upgrade when it delivers what it needs to perfectly
>> well? [...]
> Yep. That is the main reason why most of my customers stick to ZF1 since
> they don't see any economical reasons to switch.
>
>> One definite issue, however, is how to find developers with ZF2
>> experience. [...]
> This is a aspect I did not thought about yet. There must be some good
> reasons for it. One reason might be that there is no English ZF2 books
> out yet. There is one German ZF2 book out (mine ;-) but I really wonder
> why the English market for ZF2 books is not existing yet. So, English
> authors, where are you?
>
>> Documentation is often cited, and yet I hear about an equal number of
>> folks saying the documentation is horrible as are saying it is
>> wonderful; I think it's next to impossible to have documentation for a
>> project this size that will be universally useful. That said, we could
>> and should have better documentation -- the problem is identifying
>> what needs to be done, as well as getting enough attention diverted
>> from the fun tasks of coding to the tasks of writing docs -- which are
>> very, very different skills.
> Yep, I always hear people bashing the docs, but when I ask them, what
> should be done, I get no real suggestions. Ok, some components are
> missing IIRC and some component docs could be improved. But maybe people
> are looking for more docs on how to combine all components to a full
> featured ZF2 application. The user guide is only the first step.
>
> Regards,
>
> Ralf

I think this is the biggest problem with the docs currently. Zend
Framework is a component library *and* an MVC framework. The docs are
for the most part great for the library part, but are lacking for the
MVC part. We need the whole story of how components, service managers
and helpers all work together in an application.

For example, if I wanted to change the default doctype for my project,
and all I had were the Doctype helper docs, then the solution suggested
in the docs would simply not work. So I've now sent my first pull
request. Yay! It's a very minor addition to the doctype docs to include
a link to the view manager docs, which is where the doctype needs to be
set in an MVC app. It's this sort of linking that I feel is missing all
over the place. Each component has docs for usage in isolation, which is
important, but 90% of the time, the components are used in an Zend\Mvc
app and are pulled from one of the service managers which has its own
setup and configuration.

Another example is Zend\Escaper, which has fantastic docs... if you're
not using Zend\Mvc.
Luckily I knew view helpers existed thanks to the user guide, but
they're not in the View Helper docs, or even in the list of included
view helpers:
http://framework.zend.com/manual/2.2/en/modules/zend.view.helpers.html#included-helpers
They're also not mentioned AFAICS in the Zend\Escaper docs either.

Cheers,
David
Reply | Threaded
Open this post in threaded view
|

Re: Acceptance of ZF2

Spabby
I'm still not convinced there is that big a problem. 

You anecdotally say there is a problem Ralf, based on the fact that in a room full of existing ZF1 users, the adoption rate is very low. Anecdotally I can tell you that I had dinner with 4 other developers on Friday night, and the adoption rate was incredibly high. The statistical rate should always be compared to the demographic you are sampling. 

The point is, that Matthew's data is irrefutable, and it shows more and more people downloading ZF2 as time increases. This is exactly what you would expect. In all honesty I wouldn't expect the vast majority to port existing applications from ZF1; who wants to migrate an existing working application just because they can?

I understand the concerns, I had similar conversations with some of the core contributors over the weekend. It can be intimidating for a new years when faced with the "wall of config array" that they find in the skeleton app, but in all honesty, ZF2 has an easier learning curve than ZF1. You learn a tiny amount of concepts and they are re-used all over the framework. We could add convenience methods, or allow config to be defined in a more friendly format out-of-the-box, but we take a performance hit for that. I believe (and you can correct me if I'm wrong), it was discussed in an IRC meeting and rejected, we decided that performance was preferable over using the currently cool markup language to define configurations.

We have an amazing product, let's not make any doubt about that. And over time, more and more people will start new projects, and more and more will find and use ZF2. And if they choose other frameworks because they would like to define their configurations in YAML, or in ini files, and they will take the performance hit, then so be it.

Gary


On Tue, Jun 11, 2013 at 4:27 AM, David Muir <[hidden email]> wrote:
On 11/06/13 06:55, Ralf Eggert wrote:
Hi Matthew,

First off, from download statistics, we've seen a huge upswing in ZF2
adoption -- to the point that we now have a 60/40 split ZF2/ZF1
download rate. Yes, you read that correctly -- ZF2 is being downloaded
more than ZF1 now, and actually has been for around six months now.
This sounds great. Do you have any absolute numbers you can publish
officially?

[...] As several others have
replied in this thread as well, if they have a working ZF1 app,
there's often no compelling reason to upgrade to ZF2 -- why dump
effort into an upgrade when it delivers what it needs to perfectly
well? [...]
Yep. That is the main reason why most of my customers stick to ZF1 since
they don't see any economical reasons to switch.

One definite issue, however, is how to find developers with ZF2
experience. [...]
This is a aspect I did not thought about yet. There must be some good
reasons for it. One reason might be that there is no English ZF2 books
out yet. There is one German ZF2 book out (mine ;-) but I really wonder
why the English market for ZF2 books is not existing yet. So, English
authors, where are you?

Documentation is often cited, and yet I hear about an equal number of
folks saying the documentation is horrible as are saying it is
wonderful; I think it's next to impossible to have documentation for a
project this size that will be universally useful. That said, we could
and should have better documentation -- the problem is identifying
what needs to be done, as well as getting enough attention diverted
from the fun tasks of coding to the tasks of writing docs -- which are
very, very different skills.
Yep, I always hear people bashing the docs, but when I ask them, what
should be done, I get no real suggestions. Ok, some components are
missing IIRC and some component docs could be improved. But maybe people
are looking for more docs on how to combine all components to a full
featured ZF2 application. The user guide is only the first step.

Regards,

Ralf

I think this is the biggest problem with the docs currently. Zend Framework is a component library *and* an MVC framework. The docs are for the most part great for the library part, but are lacking for the MVC part. We need the whole story of how components, service managers and helpers all work together in an application.

For example, if I wanted to change the default doctype for my project, and all I had were the Doctype helper docs, then the solution suggested in the docs would simply not work. So I've now sent my first pull request. Yay! It's a very minor addition to the doctype docs to include a link to the view manager docs, which is where the doctype needs to be set in an MVC app. It's this sort of linking that I feel is missing all over the place. Each component has docs for usage in isolation, which is important, but 90% of the time, the components are used in an Zend\Mvc app and are pulled from one of the service managers which has its own setup and configuration.

Another example is Zend\Escaper, which has fantastic docs... if you're not using Zend\Mvc.
Luckily I knew view helpers existed thanks to the user guide, but they're not in the View Helper docs, or even in the list of included view helpers:
http://framework.zend.com/manual/2.2/en/modules/zend.view.helpers.html#included-helpers
They're also not mentioned AFAICS in the Zend\Escaper docs either.

Cheers,
David

Reply | Threaded
Open this post in threaded view
|

Re: Acceptance of ZF2

Sascha-Oliver Prolic
Hi Gary,

about yml, ini files, etc you are wrong. You can easily use any configuration for your application module and cache the merged config, no performance impact is there.

Best Regards

Sascha-Oliver Prolic


2013/6/11 Gary Hockin <[hidden email]>
I'm still not convinced there is that big a problem. 

You anecdotally say there is a problem Ralf, based on the fact that in a room full of existing ZF1 users, the adoption rate is very low. Anecdotally I can tell you that I had dinner with 4 other developers on Friday night, and the adoption rate was incredibly high. The statistical rate should always be compared to the demographic you are sampling. 

The point is, that Matthew's data is irrefutable, and it shows more and more people downloading ZF2 as time increases. This is exactly what you would expect. In all honesty I wouldn't expect the vast majority to port existing applications from ZF1; who wants to migrate an existing working application just because they can?

I understand the concerns, I had similar conversations with some of the core contributors over the weekend. It can be intimidating for a new years when faced with the "wall of config array" that they find in the skeleton app, but in all honesty, ZF2 has an easier learning curve than ZF1. You learn a tiny amount of concepts and they are re-used all over the framework. We could add convenience methods, or allow config to be defined in a more friendly format out-of-the-box, but we take a performance hit for that. I believe (and you can correct me if I'm wrong), it was discussed in an IRC meeting and rejected, we decided that performance was preferable over using the currently cool markup language to define configurations.

We have an amazing product, let's not make any doubt about that. And over time, more and more people will start new projects, and more and more will find and use ZF2. And if they choose other frameworks because they would like to define their configurations in YAML, or in ini files, and they will take the performance hit, then so be it.

Gary


On Tue, Jun 11, 2013 at 4:27 AM, David Muir <[hidden email]> wrote:
On 11/06/13 06:55, Ralf Eggert wrote:
Hi Matthew,

First off, from download statistics, we've seen a huge upswing in ZF2
adoption -- to the point that we now have a 60/40 split ZF2/ZF1
download rate. Yes, you read that correctly -- ZF2 is being downloaded
more than ZF1 now, and actually has been for around six months now.
This sounds great. Do you have any absolute numbers you can publish
officially?

[...] As several others have
replied in this thread as well, if they have a working ZF1 app,
there's often no compelling reason to upgrade to ZF2 -- why dump
effort into an upgrade when it delivers what it needs to perfectly
well? [...]
Yep. That is the main reason why most of my customers stick to ZF1 since
they don't see any economical reasons to switch.

One definite issue, however, is how to find developers with ZF2
experience. [...]
This is a aspect I did not thought about yet. There must be some good
reasons for it. One reason might be that there is no English ZF2 books
out yet. There is one German ZF2 book out (mine ;-) but I really wonder
why the English market for ZF2 books is not existing yet. So, English
authors, where are you?

Documentation is often cited, and yet I hear about an equal number of
folks saying the documentation is horrible as are saying it is
wonderful; I think it's next to impossible to have documentation for a
project this size that will be universally useful. That said, we could
and should have better documentation -- the problem is identifying
what needs to be done, as well as getting enough attention diverted
from the fun tasks of coding to the tasks of writing docs -- which are
very, very different skills.
Yep, I always hear people bashing the docs, but when I ask them, what
should be done, I get no real suggestions. Ok, some components are
missing IIRC and some component docs could be improved. But maybe people
are looking for more docs on how to combine all components to a full
featured ZF2 application. The user guide is only the first step.

Regards,

Ralf

I think this is the biggest problem with the docs currently. Zend Framework is a component library *and* an MVC framework. The docs are for the most part great for the library part, but are lacking for the MVC part. We need the whole story of how components, service managers and helpers all work together in an application.

For example, if I wanted to change the default doctype for my project, and all I had were the Doctype helper docs, then the solution suggested in the docs would simply not work. So I've now sent my first pull request. Yay! It's a very minor addition to the doctype docs to include a link to the view manager docs, which is where the doctype needs to be set in an MVC app. It's this sort of linking that I feel is missing all over the place. Each component has docs for usage in isolation, which is important, but 90% of the time, the components are used in an Zend\Mvc app and are pulled from one of the service managers which has its own setup and configuration.

Another example is Zend\Escaper, which has fantastic docs... if you're not using Zend\Mvc.
Luckily I knew view helpers existed thanks to the user guide, but they're not in the View Helper docs, or even in the list of included view helpers:
http://framework.zend.com/manual/2.2/en/modules/zend.view.helpers.html#included-helpers
They're also not mentioned AFAICS in the Zend\Escaper docs either.

Cheers,
David




--
Sascha-Oliver Prolic
Reply | Threaded
Open this post in threaded view
|

Re: Acceptance of ZF2

localheinz
In reply to this post by David Muir-2
I think this is the biggest problem with the docs currently. Zend Framework is a component library *and* an MVC framework. The docs are for the most part great for the library part, but are lacking for the MVC part. We need the whole story of how components, service managers and helpers all work together in an application.

I've said it before in older threads: it would be nice if contributors were required to provide 

* documentation for components newly introduced
* updated documentation for components improved

I understand it might be difficult for some developers to provide documentation because they are developers, not technical writers. But, who should provide the documentation, if not them?


Best regards,

Andreas
Reply | Threaded
Open this post in threaded view
|

Re: Acceptance of ZF2

Bas Kamer
In reply to this post by weierophinney
Matthew, Ralf and others,


On 10 jun. 2013, at 16:39, Matthew Weier O'Phinney <[hidden email]> wrote:

Documentation is often cited, and yet I hear about an equal number of
folks saying the documentation is horrible as are saying it is 
wonderful; 

On 10 jun. 2013, at 22:55, Ralf Eggert <[hidden email]> wrote:

Yep, I always hear people bashing the docs, but when I ask them, what
should be done, I get no real suggestions. Ok, some components are
missing IIRC and some component docs could be improved. But maybe people
are looking for more docs on how to combine all components to a full
featured ZF2 application. The user guide is only the first step.


It cited because, there ís an issue with documentation and yet it is very good. It's hard to define exactly what -just- not right with it… I'll suggest an approach...


"From richer interaction, better content will follow"… 

IMHO we need better tools for the documentation to invite people to share, engage, update, request, inform, etc… (endless list)

One such an idea. Please take it as an example explaining the above;

Within docs visitors have the ability to modify/add via git pull request. But there is a problem with that. Whenever I am looking to fix a problem I am not so interested in dealing with yet another problem… I would report an issue if it were really easy… Github issues however pull the issues out of context.

- I would like the ability to see inline that there are issues with the line/paragraph/page. "Indeed, someone found this issue earlier and I therefore am correct in assuming this or that". A simple "there are issues, with this line/paragraph/content", linking to github issues would help.
- I would like the ability to report inline "Hé, this (line/paragraph) links to the old ... page, but that isn't updated anymore. I think that should be removed and this should point to …".
 
Additionally a sense of uncertainty about the nature of the problem can exist. What approach should I take (if any) to solve it. Will someone hear me? Is this problem a mistake, but what is the actual concept behind this? Did this concept or policy has been changed recently. A really engaged user will know these things, but often decision are buried deep in ML or IRC-logs (where are those today btw?)

This can leave people feeling lost and frustrated which could turn into being mean to us calling zf a piece of … OR people want to help but they just need some guidance and reassurance work done won't be useless. 
- A checkbox stating "Confirm the problem to me and i'l find some time to fix it".

call it a 'counter bell post-it' feature.


I'm not saying these things are easy or there is just one way to do them…  Some of you will say - "we have this already with github issues". My point is this is just an example, to get better adoption we more involvement will be needed and i believe involvement can be invited by richer interaction.

my two cent for the day :-)

thanks
Reply | Threaded
Open this post in threaded view
|

Re: Acceptance of ZF2

Tim Fountain-2
In reply to this post by Tim Fountain-2


On 10 June 2013 18:54, Tim Fountain <[hidden email]> wrote:
On 10 June 2013 15:30, Matthew Weier O'Phinney <[hidden email]> wrote:
On Sun, Jun 9, 2013 at 6:08 AM, Tim Fountain <[hidden email]> wrote:
> If you search for anything ZF2 related on Google, results that point at the
> official site mostly point at version 2.0 of the documentation (or sometimes
> 1.12). This is bad, and is going to become more of a problem over time.
[...] 
 
I've just been looking into how other frameworks handle this, particularly those that have been around for a few years. Most seem to have version numbers in the URL, and many also have a 'latest' or 'dev' one, but they vary as to which one they link to primarily.

The solution I like best is Django's. Their docs work in the same way as ours (version number in the URL, website links to the latest released version). But they have an XML sitemap which uses the 'priority' field to influence which one the search engines link to in the listings: high priority for the 'current' version, low priority for the older ones.

I've just submitted a pull request (105 on zf-web) that adds this to the ZF site. This is my first PR, so I hope I followed the process correctly. It's also my first time using Zend Navigation in ZF2 (although I have used it in ZF1), so any feedback on ways to improve the implementation are appreciated.

All being well, with this sitemap in place, search engines would mostly link to the 2.2 docs, unless 1.12 was more appropriate. And they'd only link to the other versions if the search query was really specific to them (e.g. "ZF 2.1 view helpers").

--
Tim Fountain
http://tfountain.co.uk/
12