Adam Sroka | 1 Mar 01:06 2009
Picon

Re: Re: Great video on Craftmanship and ethics by Robert Martin

On Sat, Feb 28, 2009 at 2:30 PM, David Carlton <carlton <at> bactrian.org> wrote:
> On Fri, 27 Feb 2009 21:02:09 -0800, Adam Sroka <adam.sroka <at> gmail.com> said:
>
>> P.S. moving forward is almost always better than moving backward.
>
> For what it's worth, that's not my experience. If I'm having a hard
> time getting to a green bar, it usually means that I made a stupid
> mistake (e.g. incorrectly implemented a core refactoring), or took a
> surprisingly inefficient path towards my goal and have learned
> something over the course of that path. In both situations, rolling
> back works well for me.
>

FWIW, I did say "almost always." ;-)

The same day that I wrote that I had a situation where I did rollback.
My pair had goaded me into making a useful change. However, I was
distracted and I made the change without writing a test first. It
broke a few tests. I forged ahead for a couple minutes trying to get
things back into working order, but quickly realized my mistake. So, I
rolled back to the last subversion check-in, wrote a test for the new
thing that we wanted to add, and got all the tests to pass. Then we
switched pairs and moved on to something else.

OTOH, I have also been in situations where I realized that something
was wrong even when the tests were passing. In that situation I
wouldn't rollback. IMO, it makes more sense to go ahead and write a
test for the newly discovered thing and see if I can get it to pass. I
assumed that this was the situation Kent was describing based on his
first description. Although, his last email makes it sound more akin
(Continue reading)

Chris Wheeler | 1 Mar 01:36 2009
Picon

Re: Re: Great video on Craftmanship and ethics by Robert Martin

On Sat, Feb 28, 2009 at 2:02 AM, kentb <kentb <at> earthlink.net> wrote:

> John,
>
> The price for JUnit Max is starting at $2/month during the beta period. I
> hope this is less of a barrier than $2500.
>

This is interesting. Do tests exist in the cloud or are they local? If they
exist in the cloud I understand how you could go to a subscription model. If
hosted locally, and run on a number of workstations, how do you implement
the subscription model? Is there a connection made between JUnit Max and a
license server?

Very curious,

Chris.

[Non-text portions of this message have been removed]

David Carlton | 1 Mar 16:27 2009

Re: Re: Great video on Craftmanship and ethics by Robert Martin

On Sat, 28 Feb 2009 16:06:05 -0800, Adam Sroka <adam.sroka <at> gmail.com> said:

> OTOH, I have also been in situations where I realized that something
> was wrong even when the tests were passing. In that situation I
> wouldn't rollback. IMO, it makes more sense to go ahead and write a
> test for the newly discovered thing and see if I can get it to pass.

Yeah, I can definitely see myself coming down on that side in a
situation like that, too.

David Carlton
carlton <at> bactrian.org

Shlomi Fish | 2 Mar 11:06 2009
Picon

Recommendation for a Wiki About Software Engineering

Hi all!

In this recently-published essay of mine:

http://www.shlomifish.org/philosophy/computers/optimizing-code-for-speed/

I made a call for a catalogue of code optimisations similar to the catalog of 
refactorings in Martin Fowler's book or the Portland Patterns' Repository. The 
best way to collaborate on such a catalogue is probably using a wiki, so I'm 
looking for a good software engineering wiki to work on it.

I'm ruling out http://c2.com/cgi/wiki because it uses a primitive and horrible 
syntax, and its page titles cannot contain spaces (or anything besides letters 
which results in CamelCase). I'd prefer a MediaWiki or something more modern 
than the original WikiWikiWeb.

Regards,

	Shlomi Fish

--

-- 
-----------------------------------------------------------------
Shlomi Fish       http://www.shlomifish.org/
Interview with Ben Collins-Sussman - http://xrl.us/bjn8s

<mauke>    I'm not interested in what you're doing; what are you trying to 
           achieve?
<PerlJam>  mauke: I'm trying to achieve world peace and this regex is
           the last thing standing in my way! ;)

(Continue reading)

kentb | 2 Mar 21:02 2009
Picon
Picon

RE: Re: Great video on Craftmanship and ethics by Robert Martin

Chris,

Currently the test history is stored locally. There are many advantages to
storing it in the cloud (or on an inside-the-firewall cloudlet), though, so
I expect to go that direction.

As for license enforcement, I am relying on the honor system. I tried to
make the price low enough that if someone was going to pay, they would pay.
My reasoning is that time spent on license enforcement at the moment
wouldn't have a big financial payback and would reduce the overall chance of
product success by adding hassles and taking time away from useful features.
If the product is successful, I have several options--adding license
enforcement, relying on enterprise licenses, or making the client free and
charging for the cloud service.

Regards,

Kent Beck
Three Rivers Institute

  _____  

From: extremeprogramming <at> yahoogroups.com
[mailto:extremeprogramming <at> yahoogroups.com] On Behalf Of Chris Wheeler
Sent: Saturday, February 28, 2009 4:37 PM
To: extremeprogramming <at> yahoogroups.com
Subject: Re: [XP] Re: Great video on Craftmanship and ethics by Robert
Martin

On Sat, Feb 28, 2009 at 2:02 AM, kentb <kentb <at> earthlink.
(Continue reading)

Adam Sroka | 2 Mar 23:52 2009
Picon

Re: Recommendation for a Wiki About Software Engineering

On Mon, Mar 2, 2009 at 2:06 AM, Shlomi Fish <shlomif <at> iglu.org.il> wrote:
> Hi all!
>
> In this recently-published essay of mine:
>
> http://www.shlomifish.org/philosophy/computers/optimizing-code-for-speed/
>
> I made a call for a catalogue of code optimisations similar to the catalog
> of
> refactorings in Martin Fowler's book or the Portland Patterns' Repository.
> The
> best way to collaborate on such a catalogue is probably using a wiki, so I'm
> looking for a good software engineering wiki to work on it.
>
> I'm ruling out http://c2.com/cgi/wiki because it uses a primitive and
> horrible
> syntax, and its page titles cannot contain spaces (or anything besides
> letters
> which results in CamelCase). I'd prefer a MediaWiki or something more modern
> than the original WikiWikiWeb.
>

Well, the advantage of Ward's Wiki is the community associated with it
more than the tool (Which is intentionally quite primitive.) If you go
somewhere else you lose the access to that community. So, it's a
question of what is more important to you.

floehopper | 3 Mar 13:15 2009
Picon

Optional Scope Contract in UK

Has anyone in the UK successfully used an Optional Scope Contract (as
described by Kent Beck & Dave Cleal [1]) or similar? If yes, do you
have a sample contract I could look at?

Many thanks, James.
http://blog.floehopper.org/

[1] http://is.gd/lsOt

tony_t_tubbs | 3 Mar 14:56 2009
Picon

Re: Great video on Craftmanship and ethics by Robert Martin

--- In extremeprogramming <at> yahoogroups.com, John Roth <JohnRoth1 <at> ...>
wrote:
>
> 
> tony_t_tubbs said:
> 
> > I enjoy Uncle Bob's writings and lectures, and am coming from this
> > with the assumption I should listen to him.  However, I have
questions.  
> 
> > He talked about writing a login page, and putting all the code in the
> > GUI to start, then pulling out the business layer, then from there
> > pulling out the DB layer.  Doing what he called a spike or thin layer
> > of functionality top-to-bottom.  In theory I get it, it makes sense,
> > but in practice building testable GUIs is hard, that's why we have
> > patterns like Presenter First or Presentation Model.  If you are not
> > starting with such a pattern that enables testing, how are you writing
> > tests?  How are you mixing whatever those tests are with writing code
> > all in the 30 sec bits?
> 
> You aren't. In XP, the term "spike" means a piece of code you write,
> possibly without tests, to explore something. It's intended to be thrown
> away, and that's --almost-- what he did by pulling the data base and
> then the business logic out of the UI code. When he pulled that logic
> out is when he wrote the tests for the persistance and domain layers.
> 
> What's left is the very skeletal UI required to run the screen. That's
> very, very difficult to write tests for, which is why we try to make it
> as small as possible. All the testing that's needed is a little stuff to
> make sure the controls are connected. That can use an external test
(Continue reading)

Rhydian Thomas | 4 Mar 11:43 2009
Picon

User-stories for components

When your developing an application you can describe its features in  
terms of users performing actions to achieve some business goal, with  
a user-story ("As as X, I want Y, so that Z", for instance). I've yet  
to see this work for components though. Do you use user-stories to  
capture the requirements of a component? For example - I'm working on  
component part of a reporting system; it's a batch process which, on a  
defined schedule, extracts records from a database, performs some  
calculations and then posts the results to a web service. I've been  
wondering how I might capture the requirements of such a component and  
whether it makes sense to do so using stories?

Adrian Howard | 4 Mar 12:59 2009

Re: User-stories for components


On 4 Mar 2009, at 10:43, Rhydian Thomas wrote:

> When your developing an application you can describe its features in
> terms of users performing actions to achieve some business goal, with
> a user-story ("As as X, I want Y, so that Z", for instance). I've yet
> to see this work for components though. Do you use user-stories to
> capture the requirements of a component? For example - I'm working on
> component part of a reporting system; it's a batch process which, on a
> defined schedule, extracts records from a database, performs some
> calculations and then posts the results to a web service. I've been
> wondering how I might capture the requirements of such a component and
> whether it makes sense to do so using stories?

As a (whoever uses the reports) I need to (view such-and-such report  
on a web page) so I can (why they need the report) ?

Adrian


Gmane