[xmlWiki-developers] Wednesday evening meeting

[Radu D. <rd...@ly...>]

Hi, team.

Tonight we had our first meeting. I(Radu),Abe and Hugo participated in
the meeting. We discussed various topics , but the main one was "The System Metaphor". Below I include the conference content. As you will see, we decided for another meeting on Thursday evening 8PM CST.

See you all tomorrow.

Radu

====================================================================

[19:44]   elhugo is now Available ()
[19:44]   kalafutj is now Available ()
[19:44]   rdram is now Available ()
<rdram> Hi,guys.
<elhugo> Hi
<elhugo> jim is still not here. I wonder aboout every body else...
[19:51]   abe is now Available ()
<abe> Hi everyone. Radu, please ignore my chat invitation
<abe> Radu, did you get many replies to your email? my 
mailbox has been very quiet, recently...
<rdram> Hi, Hugo & Abe.
<rdram> You mean messages in xmlWiki-developers list? Yes, 
I got many of them.
<abe> so people should be joining shortly...
<elhugo> Are you guys getting a copy of the message that you 
send to the list? I think I can fix that in the admin tools so that 
you only get the posts from others but not your own.
<abe> I still need to figure out how to use it. Any hints?
<rdram> Yes, I got the message.
<elhugo> What do you mean?
<elhugo> abe?
<abe> how do I get access to the list (xmlWiki-developers)?
<elhugo> I am going out for a smoke... be back in ten (time 
8:57 now)
<abe> Radu, have you put anymore thougths into the project?
<rdram> Abe, you can reply to all or to 
xml...@li....
<elhugo> just send an email to 
xml...@li... and you talk to 
everyone
<rdram> Not really. NOt because I don't want to, but just 
because I would like to have the first meeting with Hugo as 
starting point.
<abe> on your previous comment, the uiuc webmail does not 
seem to recognize the 
'xml...@li...' address, that why I 
am confused
<rdram> I am using my mailcity(lycos) account and I don't have 
problems. I don't know about uiuc account. I have problems 
connecting to students.uiuc.edu using DSL and I prefer to use 
mailcity.
<abe> the webmail seems to be in its early stages...
<rdram> How do you connect to uiuc?
<abe> https://webmail.uiuc.edu 
<rdram> I haven't tried it yet, but I will. Isn't it easier for you to 
have a yahoo or lycos account? You get use it anywhere with 
no restriction. Just an opinion.
<abe> I started looking into other implementations of wikis, 
people are using scripting on the server side (cgi, php, asp...) 
not Java, I wonder how this will work?
<elhugo> I am back
<abe> I have an yahoo account, a work account and uiuc 
account, which I was planning to use only for I2CS activites, 
thanks for the suggestion.
<abe> Hugo, I was trying to get some technical talk going 
check the question before you said 'I am back'
<elhugo> On the emai... if any of you switch accounts then 
make sure you change it over at the xmlwiki-developers list. 
Just go to the project page on sourceforge and click on list.
<abe> will try to remember :-)
<elhugo> There are some implementation in Java. Hold on one 
min and I will get the url that lists all known implementations...
<rdram> Hugo, you seem to be in the Linux world. Do you see 
any real advantage of Java technologies over Microsoft web 
stuff?
<abe> there are tons of projects in sourceForge, search for 
keyword 'wiki'
<abe> radu, java runs anywhere unlike MS... (but you have to 
test everywhere!)
<rdram> Agree.
<elhugo> Ok... this is the master list of ALL known wikis --- 
http://www.c2.com/cgi/wiki?WikiEngines 
<elhugo> Microsoft is the evil empire, need I say more ;)
<abe> thanks for the link, this will be very helpful...
<elhugo> Seriously, the web was built using Un*x and then 
Linux. That other company focus has always been the personal 
user and not the server or net per se. It is only now that they 
have infiltrated the web with their .NET stuff.
<abe> Hugo, I started to read more articles from the 
OpenSource book, it is getting really interesting and difficult to 
stop reading... I can see your point on the evil empire ;)
<elhugo> From the programming point of view, it is so much 
easier to program in Linux or Un*x becuase the tools are better 
and the OS is geared toward it. I cannot live without a 
command line and a true shell.
<rdram> Hugo, this is a question that came up yesterday: are 
we gonna deploy the Wiki on a specific web container(Tomcat, 
for example)?
<rdram> Hope this question makes sense.
<abe> IMO, the ease of development is proportional to how 
used on is to an environment...
<elhugo> rdram - if we stick to Java then it really doesn't matter 
which server we use. Weblogic, Tomcat, Websphere all adhere 
to the J2EE specs and if you can deploy in one then you can 
deploy in all. I would recommend Tomcat because it is free, 
easy to set up and it is the reference implementation for J2EE.
* abe has changed the subject to: testing the subject line
* elhugo has changed the subject to: refactoring the subject line
<abe> i jus tfpound that the subject line is useless
<abe> oops,found
<elhugo> hmmmm.... maybe
<rdram> Good point. Hugo, since I live most of my time in the 
Windows world,it seems Unix tools are not that easy to use. 
What do you mean by better tools?
* elhugo has changed the subject to: Linux
<abe> in the end one either sticks to J2EE or .NET (COM or 
ActiveX,. whatever MS marketing decides to call their 
implementation)
<elhugo> Setting up a development environment is easier. The 
dir structure is standard, the commands are standard (just try 
to grep in MS). Dealing with the other aspect of code is also 
easier. CVS, secure shell, etc. Almost all of the good IDE's are 
implemented for Linux and all servers are too.
<rdram> I am interested inthe source control aspect. Do you 
think CVS is easier to use than SourceSafe?
<elhugo> I can have for virtual screen each with 1200x1600 res 
and in effect have four diffent environments with one monitor. 
Gnome or Kde are better than Windows. As a matter of fact 
Windows is a simplified version of KDE.
* elhugo has changed the subject to: System Metaphor
<elhugo> Are you guys clear on what is a System Metaphor?
<abe> on IDE's do you have a particular implementation in 
mind for our development work? what about documentation? 
plain ascii text files?
<rdram> That's a starting point!
<elhugo> What kind of starting point?
* elhugo has changed the subject to: CVS
<rdram> For our discussion. SM is the main topic, right? SM is 
a overall idea on the project.
<abe> another way to call a simple system architecture 
doument ..., I hate the fact that this industry creates more 
jargon than one can keep up
<elhugo> We are pretty much stuck to CVS since it is what is 
supported in SourceForge.
* elhugo has changed the subject to: System Metaphor
<abe> oops document
<abe> be back soon
<elhugo> Well this "simple architecture document" does not 
have any mention about technology. This is a very important 
point.
<rdram> Could you please explain more?
<abe> I agree the solution of the problem is generic. It is just 
like a server provides such service. All I care is that the service 
will exist in some form CORBA, COM, HTTP whatever...
<abe> Or simply put it is just a bunch of UML diagrams.
<abe> But for starters, a concrete representation of the system 
is very helpful. Since it is so difficult to abstract all the 
concepts whithout having done it at least once.
<elhugo> A System Metaphor is as you correctly stated a 
starting point. We start describing the system in a very high 
level. It is usually the task of the Customer to write it down with 
the help of the Developers. It is never longer than a couple of 
paragraphs and if it is runs to more than a page then it is TOO 
long. The idea is the everybody is in the same mental frame. 
The Customer has a good idea of what is expected of the 
system and the developers have the words necesary to start 
thinking about the object that they will create and code. The 
System Metaphor is also used by some to specify the bounds 
of the funcionality of th system. AND this is the important part 
to stress ---> FUNCTIONALITY.
<elhugo> No no UML yet that comes later in the CRC session 
when the implementaion plan is done.
<elhugo> We work from the general to the specific. Not the 
other way around. We become reductionist but only gradually. 
The last reduction being a class or method
<abe> Agree. the idea was to show tath the solution is 
independent of the technology.
<elhugo> Abe - correct
<elhugo> BUT.... in Open Source we run into a little Catch-22 
since most project work in a defined set of programatic 
constructs. For example, here we have stated that we are using 
Java and etc... In a non Open Source project the technology 
does not come first.
<abe> System Metaphor is a Functional Specification 
document, not the architecute document. The Functional Spec 
is visible to the customer. Architecture document for the 
development team. However, if the customer is a knowledgable 
customer then he will care for the Architecture.
<abe> On Open Source, if we feel that we can solve the 
problem in a better way, would we not drop the initially 
proposed technology for a better and improved one?
<abe> radu, are you still awake and alive?
<elhugo> Yes and no. The equivalent of the Functional Spec 
would be the System Metaphor + User Stories + Acceptance 
Test. Outside of XP the Func Spec is limited in scope in that it 
become a document for the Customer while in XP we the +++ 
that I mentioned throught the project and actually change them 
(refactor continously)
<abe> bingo!
<elhugo> I have seen (and written) too many Func Spec that 
got collected in the circular file holder.
<rdram> Abe, I am with you. Just listening. For now.
<elhugo> Were in the learning of the process are you guys 
right now in the course?
<abe> lwearning by fire...
<abe> XP really takes a different view of the problem and tries 
to break the disconnect of the traditional methodologies. And 
that is why it is so beautiful. I am also reading XP explained, by 
Beck
<rdram> Hugo - we just had "Requirements" and "Use cases" 
sessions and started the analysis.
<elhugo> On Open Source, if we feel that we can solve the 
problem in a better way, would we not drop the initially 
proposed 
technology for a better and improved one? 
Yes... if the collective agrees but usually there is a project 
dictator, for example Linus with Linux, who maintain control of 
the direction of technology and implementation.
<rdram> Hugo, how do we deal with the technology risks? In 
XP, there is no official approach for risks or I could say there is 
a reactive approach.
<abe> On Open Source, I agree with you. But in Open Source, 
the dictator also thanks the contributors many times...
<elhugo> I strongly recommend that we all read XP Installed. It 
a very practical guide while Kent's book is more philosophical. 
Also Planning XP has the details about how to go about making 
a good plan. This is one thing we will be doing continuosly... 
PLAN, PLAN, PLAN and then AIM, FIRE!
<elhugo> thank you, thank you, than you, ....
<abe> Radu, not reactive, but in iterations. One does not know 
the roadblocks until one tries the solutions. Spikes are exactly 
for minimizing risks.
<abe> thank you, thank you, than you, ....; what do you mean?
<abe> I got it!
<elhugo> Risk - We deal with risk continously by dealing with 
time and as abe mention spikes are the technology side of it. In 
buisiness we ameliorate the fears by continously 
communicating where we are, where we were and where we 
expect to go.
<rdram> Do spikes work later in the project?
<elhugo> There is no clear cut way off dealing with using a 
COT that gets deprecated... exept that you refactor the solution 
you are working on.
<abe> The risks are more in the business side than in the 
technology side. Personal interests, downsize, all lead to poor 
communication (information hiding, No bad news are welcome!)
<elhugo> During each Iteration you are dealing with "new" 
stories to implement that may require new technology or 
approaches and even at the level of the realese you are dong so 
too. Maybe you are working on a story that you really don't 
know how to implement... what do you do?
<rdram> But isn't the developer who chooses the stories he will 
implement?
<elhugo> Radu - sometimes you may want a story that is 
"cool" but you have never done it. I have mixed feelings about 
this aspect but it can happen.
<rdram> Sometimes I have this feeling but not for short term 
solutions.
<abe> if you do not know how to implement, you cannot 
estimate, so you have to (some XP lingo that I do not 
remember) until you can estimate properly.
<elhugo> Well, if you have the time and can complete the story 
in a reasonable time you include your learning curve in your 
estimate.
<elhugo> This inclusion of the learning curve in your estimate 
seem to be something that we will be doing a lot in this project, 
myself included.
<elhugo> OK back to System Metaphor...
<rdram> Good point, back to SM.
<elhugo> Do you want me to get some links about the System 
Metaphor and put them on the web page or do would you guys 
like to do some poking around on your own and then email me 
your result and I will complement them when I put them on the 
website?
<abe> either way works for me
<elhugo> radu?
<abe> Do not want to spoil the talk. It is 7:15 here in CA, I am 
really hungry, and we have been going for 1 hour now. Do you 
want to continue or allow the rest of the team to join the 
discussion.
<elhugo> I was thinking ending this disuction too. Can you 
guys email me your results by 5:00PM EST tomorow?
<rdram> If you already have some links, we can use them as 
starting point.
<rdram> Are we planning for another meeting tomorrow 
evening? Maybe we get more guys.
<elhugo> At 5:00 I modify the webpage and update it with 
whatever we have before a meeting tomorow hopefully with 
everybody else. Is this a plan?
<abe> Ward C. says the following on SM, 
(http://c2.com/cgi/wiki?SystemMetaphor )What 
ExtremeProgramming (XP) uses instead of a formal 
architecture. A simple shared story of how the system works. 
This story typically involves a handful of classes and patterns 
that shape the core flow of the system being built.
<rdram> Hugo, sounds good. Could you also send those links 
tonight by email? I would like to take a look before 5 pm 
tomorrow.
<abe> Hugo, what is up with Ward? Is he going to join the 
project as a new customer?
<rdram> Guys, overall a good session for me tonight. .
<elhugo> About Ward... he will not be actively our customer 
and his reasons are very good. I will cut and past the 
conversation with Ward and email it to yall.
<abe> I also enjoyed the meeting. But feel that we need to 
speed up our progress...
<elhugo> About the system metaphor def in Ward's wiki.... it 
needs updating and I will explain later or you may find out 
yourselft.
<elhugo> I think we do have to speed progrss although if you 
want to measure ourselves against the other groups I think we 
are doing OK.
<rdram> Hugo, do you think Ron Jefferies' book(XP Installed) is 
more practical than Ken Beck' s Extreme Programming 
Explained?
<elhugo> C U later tomorow and Radu see if you can find some 
links about System Metaphor so that we can include them in 
the web page.
<elhugo> Hugo, do you think Ron Jefferies' book(XP Installed) 
is more practical than Ken Beck' s Extreme Programming 
Explained? 
Yes... see previous answer.
<abe> good night guys...
<rdram> Guys, do you think we shlould email the whole 
conference to the other guys?
<rdram> Good night ABe and Bon apetite!
<elhugo> Yes I think so. Radu can you do so... by emailing to 
xmlwiki - developers?
<elhugo> Buen provecho Abe
<abe> yes. there are some very good points to be shared...
<abe> obrigado a todos (thank you all, in portuguese!)


Make a difference, help support the relief efforts in the U.S.
http://clubs.lycos.com/live/events/september11.asp