Author Topic: Alternatives to RUP (Read 15440 times)

PhilR · « **on:** September 20, 2002, 12:04:12 am »

We got to discussing alternatives to RUP in the EJB and UML thread http://www.sparxsystems.com.au/cgi-bin/yabb/YaBB.cgi?board=general&action=display&num=1031892685.

I promised to post some alternative approaches to using RUP and decided to start a new thread with it - enjoy.

Phil.

OPEN Process Framework
http://www.donald-firesmith.com
http://www.donald-firesmith.com/SiteMap.html
An ideal companion for EA. Bad news - requires a fee for commercial use. Good news - the fee is US$99 :-)
This site requires a bit of exploration before you realise just how good it is. You get a 30 day evaluation for free.

MBASE Electronic Process Guide
USC Center for Software Engineering
http://sunset.usc.edu/research/MBASE/EPG/
http://sunset.usc.edu/classes/cs577b_2001/guidelines/MBASEv2_2.pdf
Guru (but not Amigo) Barry Boehm is behind this framework. There are lots of good papers and other stuff here as well.

German Government's V-Model
http://www.informatik.uni-bremen.de/uniform/gdpa/
A more traditional appraoch but this site has the whole life-cycle methodology available online. The methodology is in the public domain.

Unified Process (The Book)
http://www.amazon.com/exec/obidos/ASIN/0201571692/lonsdalesystems
http://home.earthlink.net/~salhir/UnderstandingTheUP.PDF
The text book contains enough details for most experienced people to design their own light weight process. However, things like Project Management, Configuration Management, Quality etc are missing. These are some of the things that the 'R' in RUP adds to the UP.

Extreme Programming
http://www.xprogramming.com/
http://www.extremeprogramming.org/
http://www.objectmentor.com/publications/RUPvsXP.pdf
Probably needs no introduction... (Someone want to start a new thread on the role of EA and XP?)

Software Engineering Body of Knowledge
http://www.swebok.org/
Provides a brief but authorative description of each of the software engineering "disciplines" with extensive references.

Project Management Body of Knowledge
http://www.pmi.org/
Extensive description of the project management discipline. Used to be able to download the entire manual, now only some sections are available for download. Can purchase full manual at reaonable cost.

Software Capability Maturity Model
http://www.sei.cmu.edu/cmm/cmm.html
While it has come in for a lot of criticism recently, the CMM does provide lots of "best practice" pointers if you use it soley as a reference (achieveing accreditaion is a different matter). Various manuals can be downloaded for free.

jaimeglz · « **Reply #1 on:** September 21, 2002, 11:53:06 am »

Wow, Phil!

The references you have provided are really valuable.

I have a couple of suggestions for those who are not familiar with "methodologies", software development process, project management, and their associated frameworks: Firesmith's, Bohm's or the German V software development processes are huge, and it will simply not be practical for you to try them in your next short development cycle. Simply trying to read and assimilate the book by Jacobson and his amigos (450 pages) will be a daunting task. In short, evaluating a software development processes would be a considerable project by itself: very much worth doing, of course, but just consider the effort involved!

So, what is a modest developer or a project leader in hurry to do? Here are some alternatives:

For a general "lightweight" framework see Martin Fowlers book: UML Distilled (it was in its second edition last time I saw it). Very practical, down to earth, with short chapters on the essentials, it deals with UML, the development process and has good tips on project management. By the way, I was the lead translator for the Spanish edition, which came out as "UML gota a gota" --UML drop by drop-- because the arrogant barbarian who was in charge at Addison Wesley-Pearson went ahead and changed our suggested title without consulting with us. (Please note that I do not get any fees for any extra books sold, much less for advertising it! Besides, I make my living from system development, not from translations.)

For quick and powerful ideas on software development cycles, see Alistair Co-ckburn's web site: http://members.aol.com/aco-ckburn/ (delete hypen in aco-)specially his original contribution on the VW development model.

New ideas and works are continually springing up on the field, so I'm sure other participants in our forum will have other books and web pages to suggest; but I wanted to put in my two cents worth, because I am seeing some postings that are asking questions that are closely related to what has been discussed here and in the original RUP and EJB thread.

Jaime Gonzalez

jaimeglz · « **Reply #2 on:** September 21, 2002, 12:03:15 pm »

Hey! Something real funny happened to my last posting: Alistair's last name came out as "thingy burn's"! (Yes, you can make all sort of remarks on this one, but I will abstain for the moment!)

Does anyone know how can I get his real name (C-o-c-k-b-u-r-n, pronounced "Co-burn") into the posting?

Jaime

Note: Modified posting, and the problem (moralizing translation by YaBB) was solved. Thanks for the idea, Farfetch!

CJ · « **Reply #3 on:** September 23, 2002, 05:01:06 am »

PhilR,

Thank you very very much ... those are excellent links!

Here are a couple of links that may be of interest:

http://atlas.web.cern.ch/Atlas/GROUPS/SOFTWARE/OO/asp/ref/ref.html

http://www.rspa.com/ (The SE Resources are pretty good).

For a VERY VERY SIMPLISTIC process, take a look at GRAPPLE in Sams "Teach Yourself UML in 24 Hours". It's not a bad quick start ... but probably no more than that.

Farfetch · « **Reply #4 on:** September 24, 2002, 06:42:20 am »

You could just say the url is http://www.members.aol.com/c-ockburn/ and add "Delete the hyphen" to the end of the post?

alarsen · « **Reply #5 on:** September 25, 2002, 10:46:14 am »

I'll throw one more methodology into the mix. It's called Use Case Driven Object Modeling with UML or the ICONIX Process. It is a middle of the road between eXtreme Programming (XP) and RUP. The system is based on using just the necessary UML diagrams needed to produce class diagrams. We are looking at adopting it as a standard soon as I demonstrate how EA can be used in the whole process. It's self-proclaimed key to success is the use of Robust diagrams to bridge the gap between use cases and sequence diagrams. Their web site is www.iconixsw.com

Am interested in hearing from anyone else that has looked at or have experience with ICONIX.

javelin5 · « **Reply #6 on:** September 25, 2002, 11:40:12 am »

We are currently just starting to implement ICONIX where I am working. Hopefully we will have our whole team doing it for our next release. I personally have had good success with the process and the designs that came out of it.

Tim

javelin5 · « **Reply #7 on:** September 25, 2002, 11:43:00 am »

Here is a link to ICONIX Robustness analysis. There were a series of articles and you should be able to find them all.

http://www.sdmagazine.com/documents/s=815/sdm0103c/

Tim

PhilR · « **Reply #8 on:** September 25, 2002, 07:19:07 pm »

I agree. The ICONIX process is very good. One of the things I especially like is that you can see the whole process and its components on a single diagram. I reckon these days that, if it won't fit on a single page, then you are going to have trouble communicating it to everyone

I would still strongly encourage people to tailor even a "lightweight" methodology to their needs. Methodologies are like religions. All of them have some essential truths but sometimes their "followers" become more interested in "being right" than applying the wisdom.

Some "trivia" on robustness analysis. It was originally a part of Objectory, amigo Jacobson's methodology. It made a brief appearnce in RUP but has recently disappeared from it. I am guessing that this is because it is hard to formalise the meaning of the elements in a robustness diagram. This is at odds with the UML which has a strict meta-model.

Phil.

mikewhit · « **Reply #9 on:** November 17, 2004, 06:32:40 am »

Defeating Finbar Saunders mode for URLs - use the %xx notation, eg. replace 'c' with %63.

Finbar won't then recognise the word, and the URL will be clickable directly.

ie. http://members.aol.com/aco%63kburn/

Oscar · « **Reply #10 on:** November 17, 2004, 11:07:53 pm »

Better to use the following url

http://alistair.co%63kburn.us/

it saves you a redirect.

gluzm · « **Reply #11 on:** November 20, 2004, 05:57:58 am »

I was wondering what is then the underlaying methodology for EA?

thomaskilian · « **Reply #12 on:** November 20, 2004, 07:17:02 am »

The one you use. EA just supports you by UML 2.0 compliant modeling and also with a lot of OO based project management tools.

gluzm · « **Reply #13 on:** November 20, 2004, 10:23:18 am »

I do not think that any Metholodology can be used with EA or that EA is suitable for any Methodology.

Methodologies contradict each other, how could be EA suitable for everything.

Maybe I have had to ask, for which metodology is EA the best suitable tool? (Maybe it is question for the Sparx team, do they have underlaying metodology or are they just collecting recomendations and requests from the users and putting this together?)

potterm · « **Reply #14 on:** November 20, 2004, 11:08:55 am »

I think what Thomas means is that EA will support any methodology that uses UML2.0 as its notation. Clearly it will not support methodologies that use different notations such as DFDs.

One methodology that EA is well suited to is the ICONIX process, due to its support for robustness diagrams - but there is no underlying methodology intended with EA - it just supports UML, which is a notation, not a methodology.

The same applies to most other UML tools. For example, if you buy Rational Rose you will get a tool that supports the UML notation and nothing else. You do not get RUP - it has to be bought separately, and in any case you can use RUP with any UML tool - it doesn't have to be Rose...

Sparx Systems Forum

News: