You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@isis.apache.org by Stephen Cameron <st...@gmail.com> on 2015/06/10 12:51:01 UTC

chats demo application - work in progress

Hello Isis Users

I am working on a demo application using Apache Isis, my first efforts.

It's in Github and based on the simpleapp maven archetype, go here [1] if
you wish to take a look.

So far so good but nothing too fanastic to speak of as yet, other that Isis
living up to its NO promises for me.

I hope to get some work so I can spend a few weeks refining it alot
further. Its for a non-profit so should remain open-source, basically used
to manage registration of participants in activities.

One question, response does seem a bit slow to me, given that all tiers are
on the same machine, and one user, is this normal? I guess maybe yes as
there is no page compiling/caching?

[1] https://github.com/Stephen-Cameron-Data-Services/isis-chats

Re: chats demo application - work in progress

Posted by Stephen Cameron <st...@gmail.com>.
Cheers,

The online demo (http://mmyco.co.uk:8180/isis-onlinedemo/wicket/) is as
reponsive as my local chats demo, the former is reasonable given the server
is on the other side of the planet, but the later should fly. I wondered if
it was logging in debug mode, but your theory sounds a similar issue.



On Wed, Jun 10, 2015 at 9:06 PM, Dan Haywood <da...@haywood-associates.co.uk>
wrote:

> On 10 June 2015 at 11:51, Stephen Cameron <st...@gmail.com>
> wrote:
>
> >
> >
> > One question, response does seem a bit slow to me, given that all tiers
> are
> > on the same machine, and one user, is this normal? I guess maybe yes as
> > there is no page compiling/caching?
> >
>
> We have noticed that Isis has slowed down over the last release or so
> (especially when I compare to some screencasts I made a while back on
> earlier versions).  I do have a theory on that.
>
> Prior to 1.8.0 we had an @ActionInteraction annotation (also
> @PropertyInteraction, @CollectionInteraction), and if present these allowed
> a domain event to be fired for those annotated members.  In fact, the event
> would be fired five times, for
> hide/disable/validate/pre-execute/post-execute.  Subscribers on the event
> could/can veto the action etc as required.
>
> In 1.8.0 we deprecated those annotations and introduced
> @Action(domainEvent=...) etc.  What I've realised since is that we now
> *always* raise these events for every member, using default event classes
> if the domainEvent isn't specified.  That's an awful lot of events being
> fired.
>
> If my tests confirm that is the issue, then for 1.9.0 I'm thinking of
> extending @Action to be able to turn domainEvents off globally and then
> enable as required (ie reinstate the pre-1.8.0 behaviour).
>
> will let you know.
>
> Dan
>
>
>
>
>
>
>
> >
> > [1] https://github.com/Stephen-Cameron-Data-Services/isis-chats
> >
>

Re: chats demo application - work in progress

Posted by Stephen Cameron <st...@gmail.com>.
the restful objects 'viewer' is very nippy!

I can always build an XForms based client

On Wed, Jun 10, 2015 at 9:06 PM, Dan Haywood <da...@haywood-associates.co.uk>
wrote:

> On 10 June 2015 at 11:51, Stephen Cameron <st...@gmail.com>
> wrote:
>
> >
> >
> > One question, response does seem a bit slow to me, given that all tiers
> are
> > on the same machine, and one user, is this normal? I guess maybe yes as
> > there is no page compiling/caching?
> >
>
> We have noticed that Isis has slowed down over the last release or so
> (especially when I compare to some screencasts I made a while back on
> earlier versions).  I do have a theory on that.
>
> Prior to 1.8.0 we had an @ActionInteraction annotation (also
> @PropertyInteraction, @CollectionInteraction), and if present these allowed
> a domain event to be fired for those annotated members.  In fact, the event
> would be fired five times, for
> hide/disable/validate/pre-execute/post-execute.  Subscribers on the event
> could/can veto the action etc as required.
>
> In 1.8.0 we deprecated those annotations and introduced
> @Action(domainEvent=...) etc.  What I've realised since is that we now
> *always* raise these events for every member, using default event classes
> if the domainEvent isn't specified.  That's an awful lot of events being
> fired.
>
> If my tests confirm that is the issue, then for 1.9.0 I'm thinking of
> extending @Action to be able to turn domainEvents off globally and then
> enable as required (ie reinstate the pre-1.8.0 behaviour).
>
> will let you know.
>
> Dan
>
>
>
>
>
>
>
> >
> > [1] https://github.com/Stephen-Cameron-Data-Services/isis-chats
> >
>

Re: chats demo application - work in progress

Posted by Dan Haywood <da...@haywood-associates.co.uk>.
On 10 June 2015 at 11:51, Stephen Cameron <st...@gmail.com>
wrote:

>
>
> One question, response does seem a bit slow to me, given that all tiers are
> on the same machine, and one user, is this normal? I guess maybe yes as
> there is no page compiling/caching?
>

We have noticed that Isis has slowed down over the last release or so
(especially when I compare to some screencasts I made a while back on
earlier versions).  I do have a theory on that.

Prior to 1.8.0 we had an @ActionInteraction annotation (also
@PropertyInteraction, @CollectionInteraction), and if present these allowed
a domain event to be fired for those annotated members.  In fact, the event
would be fired five times, for
hide/disable/validate/pre-execute/post-execute.  Subscribers on the event
could/can veto the action etc as required.

In 1.8.0 we deprecated those annotations and introduced
@Action(domainEvent=...) etc.  What I've realised since is that we now
*always* raise these events for every member, using default event classes
if the domainEvent isn't specified.  That's an awful lot of events being
fired.

If my tests confirm that is the issue, then for 1.9.0 I'm thinking of
extending @Action to be able to turn domainEvents off globally and then
enable as required (ie reinstate the pre-1.8.0 behaviour).

will let you know.

Dan







>
> [1] https://github.com/Stephen-Cameron-Data-Services/isis-chats
>

Re: chats demo application - work in progress

Posted by Michael Lynch <mi...@gmail.com>.
Good to hear. Huge kudos on the project :)

Just in the process of trying to find a <insert everything awesome>
solution for an in house app.

Cheers

On Tue, Jun 16, 2015 at 4:54 PM, Dan Haywood <da...@haywood-associates.co.uk>
wrote:

> Now we have the Isis UI in reasonable state, we do intend to spend some
> time on the internals of Isis... we could simplify many areas
> significantly.  Performance will be one of those areas I'll be looking
> into.
>
> Thx
> Dan
>
> On 16 June 2015 at 07:46, Michael Lynch <mi...@gmail.com> wrote:
>
> > Stephen Cameron <st...@...> writes:
> >
> > >
> > > Hello Isis Users
> > >
> > > I am working on a demo application using Apache Isis, my first
> > efforts.
> > >
> > > It's in Github and based on the simpleapp maven archetype, go here [1]
> > if
> > > you wish to take a look.
> > >
> > > So far so good but nothing too fanastic to speak of as yet, other that
> > Isis
> > > living up to its NO promises for me.
> > >
> > > I hope to get some work so I can spend a few weeks refining it alot
> > > further. Its for a non-profit so should remain open-source, basically
> > used
> > > to manage registration of participants in activities.
> > >
> > > One question, response does seem a bit slow to me, given that all
> > tiers are
> > > on the same machine, and one user, is this normal? I guess maybe yes
> > as
> > > there is no page compiling/caching?
> > >
> > > [1] https://github.com/Stephen-Cameron-Data-Services/isis-chats
> > >
> >
> >
> > This concerns me too. I'm running the estatio app on an i3 laptop
> > (2.5ghz 2c2t), and accessing from a Haswell desktop - If I open 5 lease
> > objects at once (multi tab), the time until the last is loaded is 7
> > seconds, and the i3 is 70 - 100 percent CPU the whole time.
> >
> > I'm not a stickler for performance but this seems a little excessive for
> > a CPU capable of 6,000 mips
> >
> >
> >
>

Re: chats demo application - work in progress

Posted by Dan Haywood <da...@haywood-associates.co.uk>.
Now we have the Isis UI in reasonable state, we do intend to spend some
time on the internals of Isis... we could simplify many areas
significantly.  Performance will be one of those areas I'll be looking into.

Thx
Dan

On 16 June 2015 at 07:46, Michael Lynch <mi...@gmail.com> wrote:

> Stephen Cameron <st...@...> writes:
>
> >
> > Hello Isis Users
> >
> > I am working on a demo application using Apache Isis, my first
> efforts.
> >
> > It's in Github and based on the simpleapp maven archetype, go here [1]
> if
> > you wish to take a look.
> >
> > So far so good but nothing too fanastic to speak of as yet, other that
> Isis
> > living up to its NO promises for me.
> >
> > I hope to get some work so I can spend a few weeks refining it alot
> > further. Its for a non-profit so should remain open-source, basically
> used
> > to manage registration of participants in activities.
> >
> > One question, response does seem a bit slow to me, given that all
> tiers are
> > on the same machine, and one user, is this normal? I guess maybe yes
> as
> > there is no page compiling/caching?
> >
> > [1] https://github.com/Stephen-Cameron-Data-Services/isis-chats
> >
>
>
> This concerns me too. I'm running the estatio app on an i3 laptop
> (2.5ghz 2c2t), and accessing from a Haswell desktop - If I open 5 lease
> objects at once (multi tab), the time until the last is loaded is 7
> seconds, and the i3 is 70 - 100 percent CPU the whole time.
>
> I'm not a stickler for performance but this seems a little excessive for
> a CPU capable of 6,000 mips
>
>
>

Re: chats demo application - work in progress

Posted by Michael Lynch <mi...@gmail.com>.
Stephen Cameron <st...@...> writes:

> 
> Hello Isis Users
> 
> I am working on a demo application using Apache Isis, my first 
efforts.
> 
> It's in Github and based on the simpleapp maven archetype, go here [1] 
if
> you wish to take a look.
> 
> So far so good but nothing too fanastic to speak of as yet, other that 
Isis
> living up to its NO promises for me.
> 
> I hope to get some work so I can spend a few weeks refining it alot
> further. Its for a non-profit so should remain open-source, basically 
used
> to manage registration of participants in activities.
> 
> One question, response does seem a bit slow to me, given that all 
tiers are
> on the same machine, and one user, is this normal? I guess maybe yes 
as
> there is no page compiling/caching?
> 
> [1] https://github.com/Stephen-Cameron-Data-Services/isis-chats
> 


This concerns me too. I'm running the estatio app on an i3 laptop 
(2.5ghz 2c2t), and accessing from a Haswell desktop - If I open 5 lease 
objects at once (multi tab), the time until the last is loaded is 7 
seconds, and the i3 is 70 - 100 percent CPU the whole time. 

I'm not a stickler for performance but this seems a little excessive for 
a CPU capable of 6,000 mips