Building Together

[Justin Bozonier]

I'd like to build off of what Justin Wilcox and the rest of us
discussed at our wrap up. I'd like to see us come together as a group
and build something to solve a *big* problem while also utilizing the
same techniques we always clamor on about. I'd like to see us use this
thread to brainstorm large problems that we face.

This thread is only for problems. This is day dreaming about what we
wish worked differently (in our field or not). The bigger the problem
the better.

Once we've got this thread going fairly well I will get another thread
set up where we can respond about quick solutions to those problems.
The ones that seem nearly impossible are the ones we should be
tackling. I'll give a few examples of some that I have (and I'll post
some possible thoughts for solutions in a new thread):

* Aggregate the data of my life so smoothly that I get insights into
myself. There is too much data that I'm interested in and I don't know
how to keep track of it all. I really wish I had one place to go to
get up to date on literally, everything I cared about. Work email,
home email, all my RSS feeds, the latest Google searches on things I
care about, what my wife's schedule looks like, what's going on in my
community that may interest me, my IRC and Google Chat messages...
absolutely everything aggregated and presented in an easily digested
fashion.

* Programming language interoperability. I want to be able to use
Python for one set of tasks, then switch and use C# for another, and
then maybe some Lisp, all in the same app. Solving this problem would
truly allow us to herald in the era of polygot programming.

* Empower engineers to build rock solid business cases for better
programming practices. I need research completed, time saved, money
saved, charts, references for all data sources. All of this would be
PDF'd and ready for my boss by Monday morning.

If we all put out about three BIG challenges, we should be able to
come up with something huge that can reshape our professional and
maybe even our social lives (I swear I will have one some day!).

[MrJavaGuy]

I would like to add to the "data in my life", the ability to scan in
paper and organize (document management). And recognize things like
receipts for extra processing.

[Glenn Block]

I am not sure what we should do, but it would be interesting to see if we
can break new ground together rather than do "just another implementation"

Glenn

[Robert Ream]

This isn't real big, actually it's quite pragmatic... Working with
Visual Studio on various projects every day, I find myself wishing
there was some easy way to tag (and un-tag) sections of code that
relate to each other in concept, or use case, or maybe just what I'm
working on for the moment, or whatever I want. Bring some of the Web
2.0 information management and navigation patters to the environment.
I have heard there are other IDEs and/or text editors out there that
have done this, so this probably will be just another implementation,
but since I'm stuck using VS for now I haven't really looked into it.
The project folders vs. class view options that I'm forced to navigate
are so... 1993.

-Robert

[Jeff Tucker]

I've actually been thinking about writing a new development IDE that
would be similar to Visual Studio but much more lightweight and fast.
I have a number of issues with Visual Studio and I think an Open
Source solution to this (or maybe even a way to modify Eclipse to
support) would be good.

Either way, think back to that discussion "are we innovating or just
porting" back at the Alt.Net Seattle Open Spaces (the big one a few
months ago). I think we should innovate instead of just porting the
next Java tool. Maybe we should take a look at how we develop
software and even with all the stuff we currently have, are there
still any pain points that we could address? Is there anything that
we could improve on? Database migration and version management comes
to mind and we all know how painful working with DB builds and schema
changes can be. I would also like some refactoring tools for HTML and
CSS. I always wished that I could have safe delete and rename and
find duplicates and go to declaration for things like CSS tags. Might
be nice to refactor things like tables and divs and such as well
(eliminate redundant nested tables and divs maybe? Change fixed
positioning to liquid? Things like that). A "make this work in
firefox" button would also rule. How do these sound?

[Dave F]

I agree.

I am especially bitten by this problem when I want to navigate between
unit, integration and acceptance tests by feature (which are in
different test assemblies).

[Dave F]

For me, "Make this work in IE" would actually be more useful.

On Jun 30, 2008, at 10:50 AM, Jeff Tucker wrote:
>
> ... A "make this work in

[Robert Ream]

Another thing I was think that is not big either:

Since there were several Castle Monorail users in the group and a few
folks who were interested in learning Monorail as well, I also thought
maybe we could make some group contributions to the project in order
to help push it to v1.0, or maybe make a OS demo app. What would you
most like to see in the Castle project that is missing? Better
documentation? Samples? Screencasts? Intelisense for brail?

-Robert

[Duvall Buck]

I would be willing to participate in building a better Monorail demo
app.

I'm having to come up with a nice VS wizard for our company to load up
a startup project that we could extend from so making a generic one
will not be too much more.

Of course it would be helpful for others to be involved with this too.
My needs are targeted to Kaseya so my 'generic' demo would probably to
biased. Having others involved would help to keep it more generic.

On Jun 30, 12:54 pm, "Robert Ream" <robertr...@gmail.com> wrote:
> Another thing I was think that is not big either:
>
> Since there were several Castle Monorail users in the group and a few
> folks who were interested in learning Monorail as well, I also thought
> maybe we could make some group contributions to the project in order
> to help push it to v1.0, or maybe make a OS demo app. What would you
> most like to see in the Castle project that is missing? Better
> documentation? Samples? Screencasts? Intelisense for brail?
>
> -Robert
>

[Robert Ream]

The "Big Thing" I am thinking about these days is developing
frameworks and tooling around DDDD, particularly the framework needed
to make it easy to push state change awareness out to the domain
model. I am not aware of any Java frameworks or tools that allow for
this currently.

-Robert

[Justin Bozonier]

Robert the closest thing I know of is Flow Based Programming but it
gets very little attention as well though there is a framework for
both Java and C#.

Jeff, I like what you had to say. You actually help me spin myself
around and think of something big but small. What if we developed
tools to help make agile methods more alluring. One that popped out at
me is a TDD unit test framework that uses genetic algorithms to write
the code that you are writing tests for. The programs that pass the
most tests would live on and be bred with one another until they all
pass.

It would be awesome to write a set of 20 tests, step away from the
computer and come back to code that works (but may require some
performance optimization).

Then all devs would be excited to unit test. I would be even more
excited!

[MrJavaGuy]

I like this idea!

[Robert Ream]

Yes, Flow Based Programming is close, but what I am talking about is a
framework for building DDDD apps. We have ODBs or ORMs for DDD, what
will be needed for DDDD? I'm thinking that taking some of the concepts
that folks like Greg Young and Udi are working with, and try to build
a basic framework to support these types of systems. Greg is
publishing code with his series of blog posts on DDDD. I'm highly
interested in the meshing of messaging systems with DDD, even if I'm
not building highly distributed systems currently, the separation of
concerns regarding state transitions and reporting that Greg talks
about seems like enough of a "Big Deal" to me, that I'd like to
participate in exploring this. The idea of being able to plug into the
stream of domain state changes for continuous publication to multiple
stores or data warehouses for BI purposes seems pretty cool...

-Robert

[Justin Bozonier]

I think this warrants some more exploration... I'd like to help with
this regardless of what grand problem we end up choosing.

On Jun 30, 6:07 pm, "Robert Ream" <robertr...@gmail.com> wrote:
> Yes, Flow Based Programming is close, but what I am talking about is a
> framework for building DDDD apps. We have ODBs or ORMs for DDD, what
> will be needed for DDDD? I'm thinking that taking some of the concepts
> that folks like Greg Young and Udi are working with, and try to build
> a basic framework to support these types of systems. Greg is
> publishing code with his series of blog posts on DDDD. I'm highly
> interested in the meshing of messaging systems with DDD, even if I'm
> not building highly distributed systems currently, the separation of
> concerns regarding state transitions and reporting that Greg talks
> about seems like enough of a "Big Deal" to me, that I'd like to
> participate in exploring this. The idea of being able to plug into the
> stream of domain state changes for continuous publication to multiple
> stores or data warehouses for BI purposes seems pretty cool...
>
> -Robert
>

[Glenn Block]

Udi already has NServiceBus (should have mentioned this in our tools) which
addresses many of his concepts. We could look at building a fictitious
system (or a real one) based on NServiceBus. Both Udi and Greg Young are
good friends, I am sure they'd be happy to weigh in on the design (from afar
like over skype :) )

Glenn

-----Original Message-----
From: altnet...@googlegroups.com [mailto:altnet...@googlegroups.com]

[Glenn Block]

Spoke to Greg, he'd be down with helping / advising us from a design
perspective if we pursue this path. He thought it was a cool idea.

Glenn

-----Original Message-----
From: altnet...@googlegroups.com [mailto:altnet...@googlegroups.com]
On Behalf Of Justin Bozonier
Sent: Monday, June 30, 2008 8:02 PM
To: Seattle area Alt.Net
Subject: Re: Building Together

[Duvall Buck]

I would be able to give cycles to a DDDD project. We're going to
implement NServiceBus with the Castle stack in my day job so a project
like this would pair nicely with my deliverables.

So if this is the way the group decides to go I'm in.

[Aaron Jensen]

What do you mean by "implement NServiceBus with the Castle stack"?

Also, have you all looked at MassTransit?

[Robert Ream]

The first thing that stood out to me with Greg's DDDD is making a
state transition log/audit a central concern of the domain model,
essentially giving the transaction log of a OLTP system a higher level
of meaning by pushing it out to the aggregate root level. I am aware
of NServiceBus and MassTransit, though I haven't looked into either
very deep for lack of bandwidth lately. I was actually thinking that
building on top of one of the two frameworks would be a very good
idea, no need to roll our own framework for handling infrastructure
concerns like sub/pub etc., but from my brief look at them they didn't
appear to provide a framework for implementing a pervasive state
transition log within a domain's aggregate roots. If I am wrong about
this, well, then that's great, one less thing I'd (we'd) have to
write, but this seems like one of the major missing pieces for doing
DDDD (at least as Greg presents it) and not just message based SOA,
with long-running workflows and orchestrations, which is what it
looked like to me NServiceBus and MassTransit projects address
currently.

-Robert

[Duvall Buck]

We're moving our current application from a ASP/javascript application
to the Alt.NET world. The architecture we'll be using is composed of
NHibernate/MonoRail/Brail/ExtJS. We're doing TDD with DDD for
development. We're going to need a service bus to handle some of the
functionality. So we're using the Castle stack and NServiceBus as
infrastructure components to build out the rewrite of the application.
We're in the prototype phase of this right now with a demo deliverable
at the end of July. We're not going to add the NServiceBus to the demo
but it will be added in the next couple of months.

[Duvall Buck]

As far as MassTransit, I know that Ayende has blogged about it but
that is all at this time.

On Jul 1, 8:40 am, "Aaron Jensen" <aaronjen...@gmail.com> wrote:

[Erick Thompson]

I'm not familiar with NServiceBus (reading the website now), but why is it better than using service broker? I wrote some excellent message based applications on SB in SQL 2005, and it worked like a charm. I can see if SQL is a no-go, but I really like the design, if not the API.

 

Thanks,

Erick

[Glenn Block]

NServiceBus is for large scale distributed pub-sub based systems. It handles message routing and partitioning of subscribers. You can find out more here http://www.nservicebus.com/  and in this interview.

> > > > > > > > > > firefox" button would also rule.<br

[Justin Bozonier]

Here's another option... As has been said by many, Python's greatest
power is in its libraries. Maybe we could get 80% of the benefits just
by taking those ideas and translating them to .NET.

BTW Can we have a deadline on this discussion? I'm thinking by
Wednesday.

On Jul 1, 3:33 pm, "Glenn Block" <glenn.bl...@gmail.com> wrote:
> NServiceBus is for large scale distributed pub-sub based systems. It handles
> message routing and partitioning of subscribers. You can find out more herehttp://www.nservicebus.com/ and in this interview

> <http://mfile.akamai.com/14853/wmv/microsofttec.download.akamai.com/14...
> chEdOnline/Videos/EU_1_udahan_pfoster_FB_100.wmv> .

[Jeff Tucker]

Are we innovating or just porting in that case? ;)

I may have another opportunity if anyone is interested. I recently
had the opportunity to meet a professor from the UW (her name escapes
me at the moment but I got her card) who is doing research in building
an artificial hand. This thing is pretty cool. Anyway, she says that
she always has difficulties with getting good people to work on the
projects for a variety of reasons, funding in particular. I told her
that I'd be willing to volunteer and that I knew a huge group of
developers that might be interested as well because making an
artificial hand work as well as a real one sounds like an excessively
cool problem to work on. She seems totally surprised that anyone
would volunteer to work on a project like this but she gave me her
card and said to let her know. I personally would love to do this. I
don't have any background in bioengineering but she says that a lot of
people don't and I'd like to believe that I can pick up enough to
figure out the software components without too much difficulty.
Anyway, would anyone here be interested in working on something like
this?

Also, to the best of my knowledge, there is no open-source or
volunteer research-type projects out there; everything I've ever done
or heard of is all grad students or academics or whatnot. Maybe this
would be a chance to start something new and innovative in research.
If we lead by example perhaps others would follow and we could make a
difference in the world beyond just how to create better software. I
think this is also a good opportunity to show that our practices,
tools, and ideas are applicable to any type of project as well.

[Justin Bozonier]

Jeff this is serious awesome sauce.

You get my vote if you can make this happen. If everyone else agrees
then should we just email you our contact info so we can get the ball
rolling?

What does everyone else think?

> > > message routing and partitioning of subscribers. You can find out more herehttp://www.nservicebus.com/and in this interview

[Justin Bozonier]

BTW my last idea was porting as opposed to innovating... I figured I
had to balance out my other odd ball ideas. ;)

On Jul 7, 9:05 am, Jeff Tucker <fred.f.cho...@gmail.com> wrote:

> > > message routing and partitioning of subscribers. You can find out more herehttp://www.nservicebus.com/and in this interview

[Justin Wilcox]

There are a ton of really interesting ideas on this thread - great stuff.  Is there more than one project people might be interested in contributing to?  I saw a lot of people get excited about a DDDD project, obviously giving the doctor a hand would be nice too, and I imagine there are one or two other ideas that caught people's eye. 

 

I wonder if it would be helpful to apply some of the techniques we use in picking our open spaces sessions to help people pick the projects they want to contribute to.  One way would be for people who have proposed ideas to do a little due dilligence on them, perhaps coming up with a rough outline for the projects in terms of features, customers, people necessary, time available/required, etc. and going over those with the group.  Of course nothing they'd come up with would have to be set in stone, but it would help crystalize some of our options and hopefully prevent false starts.  After the projects have been presented, people could sign-up for the projects they're interested in and we could go from there.

 

We could reserve a time during the July meeting to go over the projects (is that kosher?) or we could do it virtually and perhaps record the meeting for people who aren't able to attend.

[Justin Bozonier]

This sounds like a good idea. I'll aggregate the different ideas and
post them into the solutions thread and the owners can post their
research there.

On Jul 7, 10:29 am, "Justin Wilcox" <justi...@gmail.com> wrote:
> There are a ton of really interesting ideas on this thread - great stuff.
> Is there more than one project people might be interested in contributing
> to?  I saw a lot of people get excited about a DDDD project, obviously
> giving the doctor a hand would be nice too, and I imagine there are one or
> two other ideas that caught people's eye.
>
> I wonder if it would be helpful to apply some of the techniques we use in
> picking our open spaces sessions to help people pick the projects they want
> to contribute to.  One way would be for people who have proposed ideas to do
> a little due dilligence on them, perhaps coming up with a rough outline for
> the projects in terms of features, customers, people necessary, time
> available/required, etc. and going over those with the group.  Of course
> nothing they'd come up with would have to be set in stone, but it would help
> crystalize some of our options and hopefully prevent false starts.  After
> the projects have been presented, people could sign-up for the projects
> they're interested in and we could go from there.
>
> We could reserve a time during the July meeting to go over the projects (is
> that kosher?) or we could do it virtually and perhaps record the meeting for
> people who aren't able to attend.
>

> On Mon, Jul 7, 2008 at 9:52 AM, Justin Bozonier <darkxant...@gmail.com>

> > more herehttp://www.nservicebus.com/andin this interview

> ...
>
> read more »

[Justin Bozonier]

Okay changed my mind. I'd like to get started and focus on churning
out results quickly. If that happens to be only research between now
and the meeting that's fine, but I'd really like to make this as lean
of a process as possible.

On Jul 7, 10:29 am, "Justin Wilcox" <justi...@gmail.com> wrote:

> There are a ton of really interesting ideas on this thread - great stuff.
> Is there more than one project people might be interested in contributing
> to?  I saw a lot of people get excited about a DDDD project, obviously
> giving the doctor a hand would be nice too, and I imagine there are one or
> two other ideas that caught people's eye.
>
> I wonder if it would be helpful to apply some of the techniques we use in
> picking our open spaces sessions to help people pick the projects they want
> to contribute to.  One way would be for people who have proposed ideas to do
> a little due dilligence on them, perhaps coming up with a rough outline for
> the projects in terms of features, customers, people necessary, time
> available/required, etc. and going over those with the group.  Of course
> nothing they'd come up with would have to be set in stone, but it would help
> crystalize some of our options and hopefully prevent false starts.  After
> the projects have been presented, people could sign-up for the projects
> they're interested in and we could go from there.
>
> We could reserve a time during the July meeting to go over the projects (is
> that kosher?) or we could do it virtually and perhaps record the meeting for
> people who aren't able to attend.
>

> On Mon, Jul 7, 2008 at 9:52 AM, Justin Bozonier <darkxant...@gmail.com>

> > more herehttp://www.nservicebus.com/andin this interview

> ...
>
> read more »

[Justin Wilcox]

Fine by me.