"As for XPO itself, we have some architectural changes in store. These will enable even better, more efficient XAF applications to be created, as well as provide advantages for other complex applications based on XPO."

Like Trevor, I would really appreciate expansion on "(XPO) ... some
architectural changes in store".

Only by doing so, can I assess whether XPO should be the basis for
certain classes of forthcoming projects this year. I am particularly
anxious to understand DevExpress's intentions in addressing the n-tier
concerns mentioned by Trevor.

Regards,
Alex Hoffman


Trevor Westerdahl wrote:
> Julian, I have been and still am a supporter for some time. Regardless
> of any answer provided, I will still remain a DevExpress subscriber...
> there are way to many incredible products from DevExpress and I
> am simply too happy with them for this to be anything other than a
> simple question of clarification.
>
> I ask that you consider my point of view. Consider that XPO, if selected
> by a customer, becomes the core of their application. I am sure you
> understand the significance of this as XPO has become the core of XAF.
> The dependency on it becomes very strong. The difference being that I
> cannot change XPO to meet my needs and DevExpress can for XAF. I ask you
> to imagine the significance of the dependency if your developers could
> NOT change XPO. There is no secret about questions related to n-tier
> support and I suspect many other current and potential users were
> expecting some specific comments related to this for the roadmap.
>
> Here are the Roadmap comments related to XPO:
>
>
> /"As for XPO itself, we have some architectural changes in store.
> These will enable even better, more efficient XAF applications to be
> created, as well as provide advantages for other complex
> applications based on XPO."/
>
>
> I am not gruff, I am not trying to cause conflict... I am going to be
> honest.
>
> XPO has not undergone any significant changes in over a year. I do
> appreciate that XPO will get "some" attention. I was really expecting
> what I believe many XPO or potential XPO users were specifically
> seeking... a definite plan to start developing XPO for n-tiers. A plan
> to make XPO objects, query criteria, and query results serializable.
> There are reasonable workarounds for XPO objects, but the XPView is
> unusable and the "Server Mode" is unusable in n-tier applications. The
> CriteriaObject is already serializable, but limitations regarding
> complex queries and multiple tables (essentially views) exist.
>
> BTW: I consider the linq support for the grids to be no benefit. Linq
> query criteria is not serializable and is a big gaping whole for n-tier
> applications. There are lots of people attempting to serialize an
> expression tree, but the whole "generics" aspect of linq really falls
> apart (I.e. not worry about the specific provider implementation). When
> serialization is involved, linq has real problems because the references
> to objects are dynamic and the expression tree can be greater in size
> than a query result. I really suspect that most accepted comments with
> the "other framework" that it is ready for n-tiers are not completely
> true and that these comments were based on surface-level
> investigations/trials. I mention this because the XPO is limited to
> 2-tiers. If n-tier users want "server-mode", the linq solution won't
> work. The interface will. It is a broader scope for what is currently
> supported, but will not meet all needs.
>
>
> Let me clear about my position and state some of the obvious: Nothing
> says DevExpress needs to make XPO ready for n-tier applications. I
> understand that. Assuming there is a day when DevExpress does decide to
> move XPO towards n-tier support (and possibly XAF with it), I am sure
> it will take time. I am sure there could not be firm dates, and I am
> sure that trials & tribulations will arise. However, the roadmap really
> does not tell me anything. Is one of the expected "advantages" support
> for n-tiers? Or... are the expected advantages really related to
> efficiency and the current version of XPO will likely be faster and
> stronger, but not necessarily support n-tiers? In my view, the
> main piece of the puzzle that needs to be addressed is the XPView with
> the overall expectation that XPO would support the serialization of XPO
> objects, complex query criteria, and query results/views (hopefully with
> paging/server mode).
>
> I know you can’t say XPO WILL support n-tiers in 2008, or even at the
> end of 2008, or even at the end of 2009 - and I am not asking for that.
> I am asking if DevExpress is actually making intended purposeful steps
> for XPO to support n-tiers during 2008? It would be nice to have an idea
> what you meant by that roadmap. It says XPO will get some form of
> support, but nothing more. Again, I could interpret that improvements
> toward n-tier support is what you meant, or I could interpret something
> else entirely. Can you clarify at all?
>
> TIA
>
>
>
>
>
>
>
>
>
>
>
>
> ------------------------------------------------------------------------
> Trevor Westerdahl - DX Squad
> BLOG: http://trevorunlocked.blogspot.com/
>
> ------------------------------------------------------------------------
> http://community.devexpress.com//forums/p/61809/209233.aspx#209233
>

Trevor has more or less stated what I would like to see in XPO for the future. I don't like the idea that DevExpress are not releasing any information regarding the future of XPO in the RoadMap for whatever reason; XPO is utilimately the future for all my applications and I want it to have more love, much more love! 

+1

My application is built completely on DX. (Aside from WCF and WF) -- anything that DX does has a direct impact on my products. Which is exactly why I chose DX. The quality. I have no regrets about this.

As Trevor said, I am one whose core is 100% XPO. The future of my application is tied to XPO. Was this a bad decision? Maybe, but probably not. However, XPO lacks in areas (n-tier) where I really need it to start to shine.

 

Trevor, et al

I'm going to have to get Oliver to respond to this in more detail.

--
Cheers, Julian

-----------------------------------------------------------
Julian M Bucknall
CTO, Developer Express, www.devexpress.com
julianb@devexpress.com

Personal blog at http://www.boyet.com
Company blog at http://community.devexpress.com/blogs/ctodx
Author of "Tomes of Delphi: Algorithms and Data Structures"
Read my articles in PCPlus every month
-----------------------------------------------------------
|

I would like to add my name to these ideas and requests.

As always, Trevor has expressed the ideas in the most eloquent and
complete way.

Just my 2 cents worth...
Marin.

Trevor Westerdahl escribió:
> Julian, I have been and still am a supporter for some time. Regardless
> of any answer provided, I will still remain a DevExpress subscriber...
> there are way to many incredible products from DevExpress and I
> am simply too happy with them for this to be anything other than a
> simple question of clarification.
>
> I ask that you consider my point of view. Consider that XPO, if selected
> by a customer, becomes the core of their application. I am sure you
> understand the significance of this as XPO has become the core of XAF.
> The dependency on it becomes very strong. The difference being that I
> cannot change XPO to meet my needs and DevExpress can for XAF. I ask you
> to imagine the significance of the dependency if your developers could
> NOT change XPO. There is no secret about questions related to n-tier
> support and I suspect many other current and potential users were
> expecting some specific comments related to this for the roadmap.
>
> Here are the Roadmap comments related to XPO:
>
>
> /"As for XPO itself, we have some architectural changes in store.
> These will enable even better, more efficient XAF applications to be
> created, as well as provide advantages for other complex
> applications based on XPO."/
>
>
> I am not gruff, I am not trying to cause conflict... I am going to be
> honest.
>
> XPO has not undergone any significant changes in over a year. I do
> appreciate that XPO will get "some" attention. I was really expecting
> what I believe many XPO or potential XPO users were specifically
> seeking... a definite plan to start developing XPO for n-tiers. A plan
> to make XPO objects, query criteria, and query results serializable.
> There are reasonable workarounds for XPO objects, but the XPView is
> unusable and the "Server Mode" is unusable in n-tier applications. The
> CriteriaObject is already serializable, but limitations regarding
> complex queries and multiple tables (essentially views) exist.
>
> BTW: I consider the linq support for the grids to be no benefit. Linq
> query criteria is not serializable and is a big gaping whole for n-tier
> applications. There are lots of people attempting to serialize an
> expression tree, but the whole "generics" aspect of linq really falls
> apart (I.e. not worry about the specific provider implementation). When
> serialization is involved, linq has real problems because the references
> to objects are dynamic and the expression tree can be greater in size
> than a query result. I really suspect that most accepted comments with
> the "other framework" that it is ready for n-tiers are not completely
> true and that these comments were based on surface-level
> investigations/trials. I mention this because the XPO is limited to
> 2-tiers. If n-tier users want "server-mode", the linq solution won't
> work. The interface will. It is a broader scope for what is currently
> supported, but will not meet all needs.
>
>
> Let me clear about my position and state some of the obvious: Nothing
> says DevExpress needs to make XPO ready for n-tier applications. I
> understand that. Assuming there is a day when DevExpress does decide to
> move XPO towards n-tier support (and possibly XAF with it), I am sure
> it will take time. I am sure there could not be firm dates, and I am
> sure that trials & tribulations will arise. However, the roadmap really
> does not tell me anything. Is one of the expected "advantages" support
> for n-tiers? Or... are the expected advantages really related to
> efficiency and the current version of XPO will likely be faster and
> stronger, but not necessarily support n-tiers? In my view, the
> main piece of the puzzle that needs to be addressed is the XPView with
> the overall expectation that XPO would support the serialization of XPO
> objects, complex query criteria, and query results/views (hopefully with
> paging/server mode).
>
> I know you can’t say XPO WILL support n-tiers in 2008, or even at the
> end of 2008, or even at the end of 2009 - and I am not asking for that.
> I am asking if DevExpress is actually making intended purposeful steps
> for XPO to support n-tiers during 2008? It would be nice to have an idea
> what you meant by that roadmap. It says XPO will get some form of
> support, but nothing more. Again, I could interpret that improvements
> toward n-tier support is what you meant, or I could interpret something
> else entirely. Can you clarify at all?
>
> TIA
>
>
>
>
>
>
>
>
>
>
>
>
> ------------------------------------------------------------------------
> Trevor Westerdahl - DX Squad
> BLOG: http://trevorunlocked.blogspot.com/
>
> ------------------------------------------------------------------------
> http://community.devexpress.com//forums/p/61809/209233.aspx#209233
>

I completly agree with this post and I am exactly in the same situation
posted by Trevor and others ...
My applications and code logic is fully xpo dependant ... for now this is
only a benefit without drawback ... I hope this will be the same in the
future ...
But lack of plans (or better stated, not a public plan) for 2008 (one year!)
is not a good notice.... hope to hear some news ..

Best regards,
LP

I'd like to bet that XPO is in a crossroad now. why? well it wont be soon when Entity Framework will be release. unless they do something to to features such as the "infamous " XPO NTier Support then i guess houston we've got a a competition.

is the current XPO Framework and infrastructure capable of  handling the over demanding change in techonology? i dont know. this issues should be tacked in and "XPO Road Map"

i am not willing (for now ) to migrate my framework outside XPO. zip-nil.but i dont dare write a single app under XPO of which i havent seen any major upgrade.

 

"but thats just me" - col. Do Little

Rollie Claro:

well it wont be soon when Entity Framework will be release. unless they do something to to features such as the "infamous " XPO NTier Support then i guess houston we've got a a competition.

LOL

Hi,

Thanks for the feedback, Trevor. The old search for the n-tier grail...
anyway, there are some simple answers to your questions. See below.

Trevor Westerdahl wrote:

> BTW: I consider the linq support for the grids to be no benefit. Linq
> query criteria is not serializable and is a big gaping whole for n-tier
> applications.

Well, what a stroke of luck that most applications in this world don't
require n-tier support! :-)

Linq support, of course, doesn't aim to make XPO available in n-tier
applications. The whole "server-mode" functionality is an entirely separate
block of functionality by now - it was extracted from functionality created
in the context of XPO, yes, but that's where the connection ends. Now we're
adding the feature of using "server-mode" with any data source that
implements LINQ compatibility, instead of requiring the use of our own
interfaces. I'm sure you'll agree that this is useful.

> I mention this because the XPO is limited to 2-tiers. If n-tier users
> want "server-mode", the linq solution won't work. The interface will. It
> is a broader scope for what is currently supported, but will not meet
> all needs.

Of course. To state the obvious, LINQ is not our technology, but we still
see the need and the opportunity to support it, as well as benefits from
doing so.

> I know you canʼt say XPO WILL support n-tiers in 2008, or even at the
> end of 2008, or even at the end of 2009 - and I am not asking for that.
> I am asking if DevExpress is actually making intended purposeful steps
> for XPO to support n-tiers during 2008? It would be nice to have an idea
> what you meant by that roadmap. It says XPO will get some form of
> support, but nothing more. Again, I could interpret that improvements
> toward n-tier support is what you meant, or I could interpret something
> else entirely. Can you clarify at all?

The architectural changes we are planning will result in the potential to
support a variety of layered application architectures, including
multi-tier ones. So this is good news for you and everybody else waiting
for this kind of functionality. The bad news is that we can't publish any
more information about the details of these plans - mainly because we don't
have it - and we can't make too many promises because our resources don't
allow us to.

Let me say that we are in the process of increasing the priority of work to
be done on XPO. This is not a simple thing to do because XPO is complex and
it requires much experience and care, so that important areas like
performance, stability, backwards compatibility don't evolve in the wrong
direction. We do appreciate the importance of the XPO layer in any typical
application, which is why we need to place special emphasis on these areas
(incidentally, this is also the main reason why we actually think it's a
good thing to see some stagnation on a certain level of functionality in
XPO, but that's a different discussion). In any case, my message is that we
are moving forward with XPO, we are moving, amongst other things, in the
direction you would like to see us moving, and I'm sure we will be able to
exceed expectations once everything comes together.

--
Regards,
Oliver Sturm
Developer Express Inc.
MVP C#

 

Thanks for the reply, Oliver.

Not knowing details is fine with me. Just knowing that XPO will continue to be developed independenty from XAF (I love XAF, don't get me wrong) and not just build to leverege new funcationality in XAF is comforting. You said all the right things to re-enforce that building my core on top of XPO was not a mistake.

Thanks!

Nate

Nate

On Mon, 11 Feb 2008 19:59:15 +0000 (UTC), "Nathan Laff" <> wrote:

>Just knowing that XPO will continue to be developed independenty
>from XAF (I love XAF, don't get me wrong) and not just build to
>leverege new funcationality in XAF is comforting.

Indeed, this decision was a by-product, if you like, of other
decisions made in our roadmap offsite. The further development of XPO
is an important part of our strategy.

--
Cheers, Julian

-----------------------------------------------------------
Julian M Bucknall
CTO, Developer Express, www.devexpress.com
julianb@devexpress.com

Personal blog at http://www.boyet.com
Company blog at http://community.devexpress.com/blogs/ctodx
Author of "Tomes of Delphi: Algorithms and Data Structures"
Read my articles in PCPlus every month
-----------------------------------------------------------
|

"Oliver Sturm (Developer Express)" wrote in message news:209547@community.devexpress.com...

Hi,

Thanks for the feedback, Trevor. The old search for the n-tier grail...
anyway, there are some simple answers to your questions. See below.

Trevor Westerdahl wrote:

> BTW: I consider the linq support for the grids to be no benefit. Linq
> query criteria is not serializable and is a big gaping whole for n-tier
> applications.

Well, what a stroke of luck that most applications in this world don't
require n-tier support! :-)

Linq support, of course, doesn't aim to make XPO available in n-tier
applications. The whole "server-mode" functionality is an entirely separate
block of functionality by now - it was extracted from functionality created
in the context of XPO, yes, but that's where the connection ends. Now we're
adding the feature of using "server-mode" with any data source that
implements LINQ compatibility, instead of requiring the use of our own
interfaces. I'm sure you'll agree that this is useful.

> I mention this because the XPO is limited to 2-tiers. If n-tier users
> want "server-mode", the linq solution won't work. The interface will. It
> is a broader scope for what is currently supported, but will not meet
> all needs.

Of course. To state the obvious, LINQ is not our technology, but we still
see the need and the opportunity to support it, as well as benefits from
doing so.

> I know you canʼt say XPO WILL support n-tiers in 2008, or even at the
> end of 2008, or even at the end of 2009 - and I am not asking for that.
> I am asking if DevExpress is actually making intended purposeful steps
> for XPO to support n-tiers during 2008? It would be nice to have an idea
> what you meant by that roadmap. It says XPO will get some form of
> support, but nothing more. Again, I could interpret that improvements
> toward n-tier support is what you meant, or I could interpret something
> else entirely. Can you clarify at all?

The architectural changes we are planning will result in the potential to
support a variety of layered application architectures, including
multi-tier ones. So this is good news for you and everybody else waiting
for this kind of functionality. The bad news is that we can't publish any
more information about the details of these plans - mainly because we don't
have it - and we can't make too many promises because our resources don't
allow us to.

Let me say that we are in the process of increasing the priority of work to
be done on XPO. This is not a simple thing to do because XPO is complex and
it requires much experience and care, so that important areas like
performance, stability, backwards compatibility don't evolve in the wrong
direction. We do appreciate the importance of the XPO layer in any typical
application, which is why we need to place special emphasis on these areas
(incidentally, this is also the main reason why we actually think it's a
good thing to see some stagnation on a certain level of functionality in
XPO, but that's a different discussion). In any case, my message is that we
are moving forward with XPO, we are moving, amongst other things, in the
direction you would like to see us moving, and I'm sure we will be able to
exceed expectations once everything comes together.

--
Regards,
Oliver Sturm
Developer Express Inc.
MVP C#

---

--
Regards,
Oliver Sturm
Developer Express Inc.
MVP C#

---

http://community.devexpress.com//forums/p/61809/209547.aspx#209547