______________________________________________________________________

							PSE NOTES
______________________________________________________________________

2001-07-04 Final Coach mtg
David Jud, Michael Locher, Adrian Kuhn, Matthias Rieger
- next year: AK, ML perhaps; DJ probably not
Conclusions:
- relationship w customer was bad
  - not always technically up to scratch
  - not always committed (eg planning games)
- University customer might be better
- final product should be *used*
- planning game should be repeated (if possible)
  - missing proper requirements collection
  - perhaps have special stories collection session?
- status reports worked well
  - tracking tasks & subtasks
- tougher reqts for testing
  - insist on JUnit (also in ESE)
  - dedicated test builder?
- code & design reviews
  - earlier
  - coaches should supervise reviews more closely
- mutual reviews can help teams overview each other
- testat requirements must be clear up front
  - all deliverables plus each team member responsibility
  * What to do about students who don't work?
- coach responsibility should be clearer
  - whom to ask for what
- wiki: notify students of important changes
- warn students about differences between ESE & PSE
- missing competition between groups
  - important for motivation!
- search for HA earlier with the right background
- resolve contract issues earlier
- more on design patterns in ESE
______________________________________________________________________

2001-06-25 Feedback session

PSE1:
- better if customer wants finished product
- contest at end to pick one project
- should have stated only prototype was expected
- good to try new technologies
- happy with coaches
- infrastructure chaotic
PSE2:
- learned a lot
- stories were motivating; negotiating with client
- open reqts were exciting (vs fixed product)
- complexity ok; quantity too much
- mixed group -- good match to technology mix
- good to have EJB/servlet project in future
- liked working on prototype rather than fixed product!
- better contact with customer would have helped
- at beginning hard to tell what exactly client wanted
  - chaotic changes in reqts
  - would have been better to have more freedom explicitly
- no worthwhile feedback from customer
- customer badly informed before meetings
- presentations on EJB and certificates very confused
- should have made code licenses clear at beginning
- each team should have mix of CS and external students
- team forming was chaotic
  * avoid changes in teams after forming
PSE3:
- real customer good for motivation
  - but problems with tools
- iteration worked well
  - XP didn't really work
- teamwork; communication ok despite problems
* expert pool would be useful
  - each HA could support some specific topic
- HA could be a stronger coach (leader)
- web site chaotic - hard to find stuff
- P2 & ESE provided good background
- missing crash course in configuration tools
PSE4:
- intimidated in beginning by new technology
- customer was not really "real"
- project was not too big
- good exposure to current technology
- needed better startup (first demo)
- iteration hardly possible without JUnit!
- XP didn't really work, but ideas were useful (planning game)
  - not really practical given different schedules
- too many status reports
* code review at end not useful
  - earlier would be better
- what to do about unmotivated team members?
- shouldn't rebuild team after initially formed
- hardest part was new technology
- coaching worked well
  - coach forced them to do things that worked out best in the end
- a hot-line would be useful for emergencies
- templates very useful
- good handouts; hard to find stuff on swiki
- ESE: should practice more testing; programming style
- team dynamics were great
PSE5:
- learned a lot from trying to use complex technology
- complexity ok
- useful to be warned before PSE starts what technologies needed
- big problems with DB
- some reqts should have been clear earlier (e.g., ANT vs Make)
- ESE needs more on time estimation
- difficulties transferring knowhow within team
- coordination difficult (finding common time slots)
PSE6:
- problems with tools
- good to have real customer
- customer did not seem motivated!
  - more a feasibility study than a real project
  - customer wasn't sufficiently technically competent
  - communication inadequate
- not really XP
  - hard to partition project to enable XP (?!)
- perhaps a smaller project would be better
PSE7:
- initial use cases too big to complete
- new technologies sapped energy from project management
- iteration difficult because new stories didn't build on old ones
- missing a project leader who could maintain overview
- coaches worked out well
  - personal contact better than in ESE
- late intro to certificates was a disaster
- code review should also focus on positive aspects!
- ESE project should be smaller?
- disappointed not to have a complete project

AK:
- EJBs motivating
- a lot of work coaching two groups!
- better prepared than last years PSE
ML:
- more complex technology than last year
- more information available on wiki, but hard to find
DJ:
- good communication within teams

______________________________________________________________________

ON FEEDBACK
- my requirements fulfilled: 
  - real, open requirements
  - new technologies; interesting problems
  - multiple iterations
  - teamwork, planning, negotiation
- problems:
  - too big; impossible to really complete project
  - customer still in exploration phase
    - interest more in experience than product
- questions:
  - better to have University-internal customer?
  - better to have naive or expert customer?
  - better to have conventional or experimental project?
  - what parts were most useful as learning experience?
  - what parts should be changed/eliminated
______________________________________________________________________

2001-06-26 Coach mtg

PSE1 (DJ): missing test archive; code archive broken
PSE2 (DJ): missing source archive
PSE3 (AK): ok
PSE4 (ML): unacceptable doc
- working on it
PSE5 (ML): build procedure broken (missing build files)
- will check this evening
PSE6 (MR): ok -- good, except for design doc
PSE7 (AK): ok
PSE8 (MR): ok -- some parts good; others uneven; bad design doc
- Mirco Rechsteiner contributed only on design doc; did lousy job
- MR will send email to make appointment
______________________________________________________________________

2001-06-19 Coach mtg

- Coaches should hammer groups to get deliverables out in time
- IGE workshop: who should attend?
- ESE HA: who would be good? (ML available)
* need to uninstall Postgres, JBOSS, ...

PSE1 (DJ): late for demo; migration demo worked well
PSE2 (DJ): well organized
PSE3 (AK): disaster -- nothing worked
PSE4 (ML): problems with Borland server (broken by group 3?)
PSE5 (ML):
PSE6 (MR): David Jrg continued implementing after Demo
- certificates now work
PSE7 (AK): good demo
PSE8 (MR): no progress; Pascal Schuppli asked for MR feedback
______________________________________________________________________

2001-06-13 Final demo

PSE 3: version problems with security concept -- can't run demo
PSE 5: 
PSE 4: nice demo; lots to show
PSE 7: 
PSE 8: demo broken?!
PSE 2: good demo; lots of tests; missing integration multiple versions
______________________________________________________________________

2001-06-12 Coach mtg

Demos at IGE: 3 parallel sessions (2 with PCs, 1 with laptops)
Feedback session: each team should prepare

PSE1 (DJ): 60% scope reached
PSE2 (DJ): integration taken out of scope; 75% chance to reach scope
* PSE3 (AK): good progress through weekend -- green status
* PSE4 (ML): some good stuff, but probably will miss some stories
- worked well; now good EJB expertise
PSE5 (ML): bad feeling; failed to do basic Beans ...
- got caught by new technology
* PSE6 (MR): well organized; but lost motivation
- renegotiated scope; drowning in technical problems
PSE7 (AK): still red; poorly motivated; small story to show
- main story will probably be vaporware
PSE8 (MR): badly organized; team not working well
______________________________________________________________________

2001-06-06 Code reviews

PSE1 Rolf Knzli (PSE4)
- lots of small technical errors
- no tests
- missing comments
- code distributed and individually reviewed

PSE2 Simon Bnzli (PSE5)
- lots of small details
- group of three reviewed all code
- reviewers were not programmers!

PSE3 Kai Rolle (PSE6)
- 4 man hours; 1200 LOC
- one bean per reviewer
- missing comments made code hard to read
- code review came too late in the course

PSE4 Mauricio Seeberger (PSE7)
- lots of mistakes
- 5 serious errors; 4 medium
- broken javadoc
- some incomprehensible methods
- no Strings -- char Arrays instead!
- very strange exception handling within methods
- each reviewed 2 classes individually
- time for review messed up planning game

PSE5 Manuel Stadelmann (PSE8)
- debug code left
- exceptions used for flow control
- three reviewers; individual
- no tests

PSE6 Stephan Zger (PSE1)
- three individual reviews
- no big problems
- hard coded URLs
- some silly (redundant) comments
- missing private declarations

PSE7 Aljoscha Bachmayer (PSE2)
- reviewed only bean classes
- good variable names
- poor documentation; hard to follow

PSE8 Michael Rogenmoser (PSE3)
- mostly small errors
- well-structured
- ~250 LOC/hr
- earlier, internal code reviews would be useful
______________________________________________________________________

2001-06-05 Coach mtg

- MR will check what MG need for user documentation (snapshots?)

PSE1 (DJ): on track
PSE2 (DJ): no progress since last demo!
- gave deadline for iteration planning
* meet with them tomorrow; will let me know what's happening
PSE3 (AK): on track
PSE4 (ML): missed status report last week (responsible missing) -- fixed
- code review done (distributed; code received late)
- database crashed, but didn't inform DJ ...
PSE5 (ML): recovered (!); fixed organization efforts
- precise iteration plan with tracking
- code review done
- demo status unclear; trying new technology far too late ...
PSE6 (MR): well-organized as usual
- renegotiating stories with customer
- tech problems with servlets
- code review should be ok
PSE7 (AK): no iteration planning; but one huge task
- unhappy about third new technology
- not very motivated but making progress
PSE8 (MR): disorganized; delivered code to review late
- will produce something; but will be weak
______________________________________________________________________

2001-05-29 Coach mtg

- AK will give testing presentation
- MR will give code review presentation
- Code review: two beans; two servlets (or equivalent)
- Code provider should propose the code to be reviewed
- Coaches vet the process

PSE1 (DJ): monitoring; seems ok
PSE2 (DJ): ditto
PSE3 (AK): planning game ok
PSE4 (ML): ok
PSE5 (ML): missed last mtg; not working <=== Serious problems?
PSE6 (MR): good progress; high performance; planning game well-prepared
PSE7 (AK): not excited by stories selected
PSE8 (MR): team not working well; planning game poorly prepared
______________________________________________________________________

2001-05-22

- ML: PSE 7 -- weak demo
	PSE 5 -- no stories completed!; focussing on security
	PSE 4 -- nice swing gui; migration succeeded; corba link to IGE
- DJ: PSE 3 -- search addresses; text only
	PSE 1 -- message between users
	PSE 2 -- hierarchical access control
- MR: PSE 6 -- nice facade with servlets; connection broken; showed test cases
	PSE 8 -- ...
______________________________________________________________________

2001-05-15 Coach mtg

Demo schedule
- PSE 7, 5 and 4 demos at Uni with Locher &  Nottaris
- all others at IGE

Review
- big code review

Database
- switch to PostGres

Troubles
- replies from PSE 1, 3 and 6

PSE1 (DJ): rescoping (problems with JBoss & interbase; switched to PG)
PSE2 (DJ): no problems?
PSE3 (AK): might miss some stories (interbase)
PSE4 (ML): realized migration problems (scoped out w customer)
	Perhaps will try anyway.
PSE5 (ML): behind on stories (2 of 4); security scoped out for now
PSE6 (MR): good progress; missing some stories
PSE7 (AK): way behind schedule (1 story of 6 ready); comm problems?!
PSE8 (MR): internal problems; missing overview of stories
______________________________________________________________________

2001-05-10 Mtg with MR

Problems with demotivation; licenses

PSE1 (DJ): demotivated; Dominic Stoll unhappy with licenses
PSE2 (DJ): ?
PSE3 (AK): not a big problem
PSE4 (ML): ok
PSE5 (ML): ok
PSE6 (MR): miscommunication grudge; late customer reactions; o/w motivated
PSE7 (AK): motivation problems due to missing security information
PSE8 (MR): phlegmatic but o/w ok
______________________________________________________________________

2001-05-01 Coach mtg

All planning games done.  Mostly success.  SB was not well prepared.
- 1 hour not enough time
- customer needs to realize importance!
- still better than last year
- groups and customer both happy in general
- cost estimates varied wildly
* should now fix customer/team mappings!
- some groups (1, 2, 5, 6, 7) did not sell migration time
______________________________________________________________________

2001-04-24 Coach mtg

All planning games at least tentatively scheduled now.

ML has set up a migration wiki page.  Has deployed his bean on IGE's server.  Now working on OpenORB -- problem with stub compiler.  NB: no support for JSP.  Teams should assign someone to this task & start migration experiments.

PSE1 (DJ):
- good demo with DB & JSP running at home
PSE2 (DJ):
- weak prototype
- poor communication -- nothing prepared till last minute
- now coding 1st prototype
PSE3 (AK):
- now DB connection working
- code review -- good results
- budget 8 hrs pppw (expandable to 20)
PSE4 (ML):
- nice Swing GUI
- DB problems; works erratically; not working yet with GUI
- have assigned CORBA task
PSE5 (ML):
- budget of 7hrs per person per week
- demo with servlets; no DB connection
- text GUI with DB
PSE6 (MR):
- minimal demo with JSP, but all pieces there (DB, tomcat ...)
- writing DB causes crashes!
- very busy
PSE7 (AK):
- now have a running prototype
- preparing planning game; evaluating use cases
- budget 8 hrs pppw
PSE8 (MR):
- nice swing demo
- preparing for planning game
______________________________________________________________________

2001-04-18 Demo 1
PSE3:
- simple swing gui to enter addresses and query them using beans
- one session bean + multiple entity beans
- *** currently no connection to DB
- tried but not working yet
- risk: migrating to IGE application server (need test AS from IGE)
- tasks distributed: (i) management, (ii) GUI, (iii, iv) coding, beans,
  (v) DB, (vi) architecture/design
- risk: still need to evaluate use cases (CRC session tomorrow?)
PSE7:
- swing gui; entity bean; session bean; text client; JBOSS
- communication w JBOSS ok, but can't get data out of entity bean yet!
  (main technical problem)
- trying with v2 of DB, but still problems
- GUI both with Swing (by hand or with builder)
  and web interface (servlets + JSP)
- still evaluating which is best
- bean management should simplify migration problems
  -- nothing JBOSS specific
- minutes taker; chief programmer; toolsmith; tester; bean provider;
  documenter; backup chief programmer/webmaster
______________________________________________________________________

2001-04-17 Coach mtg
PSE1 (DJ):
- workplan ok
- still coding the demo
- problems with JBOSS
PSE2 (DJ):
- will also use JSP (!)
- will demo on laptop
PSE3 (AK):
- great workplan
- still coding (don't know quality)
- using Swing
PSE4 (ML):
- coding
- from emails seem to be having some success
PSE5 (ML):
- coding on Saturday ...
- problems with DB
PSE6 (MR):
- staying with JSP
- seem not to be taking prototype seriously!
PSE7 (AK):
- various demos working; should be ok
PSE8 (MR):
- online workplan only excel format!
- missing status report
- swing gui is working (?)
- various demo stuff working (beans, db)

Each group now has 50MB.
______________________________________________________________________

2001-04-10 Coach mtg
PSE1 (DJ):
- first status report late
- want to use JSP! (have web designer in team)
- want to install tomcat (servlet server) -> more quota
PSE2 (DJ):
- still deciding Swing or Text GUI
PSE3 (AK):
- well-organized
- lacking strong programmers
PSE4 (ML):
- two sub-teams: GUI & server
- on track
PSE5 (ML):
- want to use JSP; develop at home
- should be able to manage (JSP guru in group?)
PSE6 (MR):
- also want to use JSP (no guru?!)
- meet on wednesdays
- have installed tomcat
PSE7 (AK):
- doing good work
- bean example on their home page!
- want to put some first security stuff in demo 1
- 3 students from WI -> also a big project
PSE8 (MR):
- chaotic impression
- now have a workflow for protocols
- scheduling problems
- clear plan to first prototype

JSP groups: 1 5 6 7(?)
- major risk -- conflicts with user reqts (MVC)
- must work in first demo!

Swing: 2(?) 3 4 8
- use netBeans (on windows) to build GUI?

ML will visit IGE to test migration path
______________________________________________________________________

2001-04-05 Housley phone call
PSE: student notebook connections to Uni net?
- Fritz Butikofer
- wireless LAN already there (need wireless cards ~200 Fr)
- or need ethernet card
- might also be network points in Pool and upstairs
- problem is to have DHCP server on same subnet
- planning connections and wireless!
- install pilot plan; security on top (later)
- would need id accounts to surf web
- DHCP to get IP address
______________________________________________________________________

2001-04-04 Class session

- IGE: Matthias Guenter, Pietro Notaris, Johannes Lang, Sabine Blaser

	Risk Analysis Presentations
PSE1 Dominic Stoll
- EJB no exp -> prototypes
- Appl server JBOSS
- multiple formats for two servers
- security certificates new tech
		[MG: could prepare tutorial; Lang responsible]
- use cases very general
- SQL -- schema unknown
- DB access
- UI reqts unknown
		[MG: need not be swing]
PSE2 Michael Zaugg
- similar problems
- missing good project planner
- mixed team (CS & non-CS)
PSE3 Renzo Killias 
- split into general, technical, financial, personnel risks
- general:
	- client commn
	- product acceptance
	- scope, design, changing reqts
- technical:
	- EJB/CORBA unknown
	- client platform unknown! -> system tests
		[MG: could make special server available through net]
	- DB -> tests
- costs:
	- testat, exams, miscalculation
- personnel:
	- teamwork
PSE4 Stefan Zapf
- main risk: resources (time)
- no one knows EJB, GUI, Interbase ...
- admin will be missing in last weeks of course!
- performance of Java?
- migration!
- specs too vague to get started
PSE5 Manuel Stadelmann
- Project risks
	- hard to evaluate costs -- minimal reqts?
	- machine resources too small
			* can students connect laptops to Uninet?! (forbidden)
	- MS has 2 weeks mil service
- Technical risks
	- EJB, DB, migration
	- java tiff import (solved?)
	- performance
	- swisskey product unknown ... (proprietary product)
PSE6 Heiner Rohner
- EJB unknown
- JSP/Servlets unknown
- JDBC unknown
- Security certificates
	- leave out in 1st phase?
- Migration
- Prog envt/tests
	- which envt? -> Unix + laptops
PSE7 Gregor Gabriel
- similar story
PSE8 Michael Mugglin
- also similar
- GUI -> early prototype
______________________________________________________________________

2001-04-03 Coach mtg
- next week Planning Game lecture by MR
- assign one IGE person per team for weekly reports & planning game
PSE1 (DJ): not well organized yet; no chief; also problems in ESE
- meetings Mondays Mensa @ 13h00
- group mtg Fridays
- chief Sascha Michel
PSE2 (DJ): well organized; tasks defined & distributed
- chief Marc Hinnen
PSE3 (AK): only two CS majors; others BW & Geography
- good start; roles defined
- chief Renzo Killias
PSE4 (ML): good start
- chief Susanne Wenger
PSE5 (ML): no homepage yet; but wiki there
- most CS and one physics student
- meet Friday @ 10h00; with ML @ 11h30
- chief Bernhard Mueller 
PSE6 (MR): not yet met
PSE7 (AK): done workplan, roles etc etc
- 5 CS + 1 BWL
- chief Divya Saxena
PSE8 (MR): mixed; good start
- meet Thursdays at 18h00
- chief?
______________________________________________________________________

2001-03-28
______________________________________________________________________

