|
Well, yes, I suppose that's obvious, but these things really need some formal definition so people know what is expected of them. Personally, I think it's too management topheavy.
Marc
Help! I'm an AI running around in someone's f*cked up universe simulator.
Sensitivity and ethnic diversity means celebrating difference, not hiding from it. - Christian Graus
Every line of code is a liability - Taka Muraoka
Microsoft deliberately adds arbitrary layers of complexity to make it difficult to deliver Windows features on non-Windows platforms--Microsoft's "Halloween files"
|
|
|
|
|
Marc Clifton wrote:
Personally, I think it's too management topheavy.
And it would be, except that one person can take more than one role. It would be nice if people could "slide" into it according to their abilities, but I don't know how that would work. Maybe in the course of discussions, we will get some kind of idea of who will do best at what.
"Blessed are the peacemakers, for they shall be called sons of God." - Jesus
"You must be the change you wish to see in the world." - Mahatma Gandhi
|
|
|
|
|
Marc Clifton wrote:
Personally, I think it's too management topheavy.
Me too 
- Anders
Money talks, but all mine ever says is "Goodbye!"
|
|
|
|
|
First off wanted to say thank you Jason. You've done a great job so far and we wouldn't be this far without your lead. ![Rose | [Rose]](https://codeproject.global.ssl.fastly.net/script/Forums/Images/rose.gif)
As for the proposed structure. I like the fact that it lays out roles. Who fills the roles should be up to the PLs for each project. While this particular structure may not be the best in the opinion of everyone, I think it is a good start.
No doubt some individuals will fill more than one role. Some roles will have more than one person. The PL will probably fill most of the lead roles.
The structure lays out an important timeline of how to proceed.
1. We gather resources (people) for each project. (PL)
2. Then divide assignments to each person based upon qualifications and strengths. (PL)
3. Then define requirements for the project. (Those interested in project, or just involved)
4. Decide what will be included in first rev. Keeping in mind that the shorter the dev cylce the more likely success. Future revs to follow. (PL and leads)
5. Requirements defined and documented (Systems Engineers)
6. Review and testing of requirements to ensure everything in order. (System Testers. Should not be same as Systems Engineers. Never debug your own work.)
7a. Once requirements approved by PL coding begins. (Software Engineers)
7b. In parrallel, Test plans are created and documented based upon requirements documentation. (Software Testers)
8. Coding completed and reviewed by testers. Modifications made if discovered. Additional tests documented based upon specific code segments.
9. Testing done and results reviewed.
10. Fixes made due to test results.
11. Repeat steps 8-10 as necessary.
12. Product release.
As for QA, this would be active throughout entire process. They would be integral in code reviews. Code reviews should be periodic and not wait till end. Final code review for purposes of forming new tests.
--
"The money power of the country will endeavor to prolong its rule by preying upon the prejudices of the people until all wealth is concentrated in a few hands and the Republic destroyed."
-- Abraham Lincoln
|
|
|
|
|
Firstly, the intent wasn't to formally structure or "compartmentalize" a team. I just wanted to formally identify roles that define a well balanced software team. As was mentioned, some people will likely fill several roles at the same time, while others may not. So for a clearer definition of the terminology, consider this:
Systems Lead - When assuming this role, you need to view things from the system-level. That is, not to have vision clouded by the details of the bug you spent all morning trying to fix. A system-level view has one target audience: the customer. What does the customer need? (Remember, the customer most likely doesn't know what they want.) How do you plan to meet that need? Is the current design in line with the project goals and mission statement? The systems lead has the highest view of the project, sometimes higher even than the project manager.
Systems Engineer - These people carry out the system-level objectives and produce concrete output in terms of documentation and system design. The systems engineer carefully considers the customer's need and approaches it with a software mind. The systems engineer takes these high-level goals and visions and breaks them down into system-level requirements. That is, functional requirements that are divided into manageable and logical components that, when all met successfully, roll up into a deliverable product and lead to a happy customer.
Software Lead - The software lead is different than the systems lead in that the level of requirements is lower. Rather than dealing with global high-level requirements, the software lead has a logical set of functional requirements from which to work. This is where the rubber meets the road so to speak. The software lead considers things like the actual implementation of a certain module of code. (Notice the systems engineer never even came in contact with the mechanics of the actual implementation.) The software lead keeps the software engineers focused and on task.
Software Engineer - These are the individuals responsible for designing the detailed mechanics of an implementation and actually coding it. In all aspects, these individuals are more involved than the proverbial "code-monkey". In this infrastructure, the software engineer commits more than just code to the project, they commit experience and design. The software engineer has a dark side as well... Just as the systems engineer documents and delivers system-level functional requirements, the software engineer is responsible for documenting the design and mechanics of the actual implementation. These design considerations will become a requirements specification in and of themselves, those of which will have been implemented either by the software engineer himself, or maybe even given to one of the aforementioned "code monkeys" to implement the detailed design.
Test Lead - I believe testing is the most crucial aspect of software development and the most expensive task to tackle for software houses. If a software project is going to skimp to meet a deadline, it's the testing that gets the axe first. With a resource pool of this size and quality (and the fact no one's getting paid here), I feel it is paramount that we not skimp on this aspect at all. In this light, the test lead is responsible for managing two teams. A software test team and a system test team. I will explain these in more detail next, but they are two separate terms. The test lead is the liaison between all the functional groups. Think of it in terms of verification and validation. First, we need to verify that all functional requirements are met, then we need to validate that the requirements are satisfactory to fulfill the needs and intentions of the system. Any anomalies found in the testing phase are to be formally documented and addressed by an anomaly committee.
Software Testers - Software testers are basically unit-testers. They test individual functions and implementations at their lowest level and throw every bad coding practice they can at the code. Their goal is to break the implementation at any cost (within engineering and common software reason of course). They reference the software engineer's documentation and look for holes and shortcomings in requirements. They are basically verifying that no stone is left unturned.
System Testers - System testers put it all together and test the software as it's intended to be used. I like to break system level testing out even further to two categories: 1. Failure-Mode testing. and 2. End-User testing. In failure-mode testing, the system tester intentionally fails (or tries to fail the system). How does the system respond? Will it be easy to figure out what went wrong from the end-user point of view? Are error messages understandable and consistent? Then end-user testing is just that. The software is being tested as it will be used by the customer. This is the "alpha" stage of development, full-up software (mostly), being tested as it's going to be used. However, don't think testers get off that easy... They have their own set of documentation as well. As the software tester tests to the software engineer's documentation, the system tester tests to the systems engineer's documentation. This is the "validation" portion. Do the requirements laid out in the system design meet the needs of the customer?
Configuration Management Lead - Ahh yes, everyone's worst nightmare.... (next to QC of course  ) This person keeps tabs on all the details. As a CM lead, you are responsible for which version of software is where, who's working on what, what anomalies are outstanding and what level of severity do they carry with them, are the people doing what they documented they are doing. etc... We'll get way more into detail on this as time goes on...
QC/et.all - These individuals are responsible for keeping tabs on everybody. They are also responsible for documenting process and insuring everyone follows that process. They also sign-off on formal software releases. (more on them later too)
Well, I hope that clears up some stuff. We'll expound on all this in more detail as we move along.
- Nitron
"Those that say a task is impossible shouldn't interrupt the ones who are doing it." - Chinese Proverb
|
|
|
|
|
That was fantastic Nitron, a wealth of knowledge and good sense.
I certainly see us putting together a CPSF (Code Project Solutions Framework) for developing software. We cannot just come out of CP2 with just applications. We should come out with a strong process model as well. From the sounds of it you would be the lead of that project
Paul Watson
Bluegrass
Cape Town, South Africa
Chris Losinger wrote:
i hate needles so much i can't even imagine allowing one near The Little Programmer
|
|
|
|
|
Paul Watson wrote:
That was fantastic Nitron, a wealth of knowledge and good sense.
... it happens every now and then. 
- Nitron
"Those that say a task is impossible shouldn't interrupt the ones who are doing it." - Chinese Proverb
|
|
|
|
|
I'm not sure I really like that structure, it needs a lot of people.
To be honest it looks like something they use at IBM or other big old compagnies 
I think the project lead should pich his own structure, because the structure always depends on the team, and the people on the team.
- Anders
Money talks, but all mine ever says is "Goodbye!"
|
|
|
|
|
Anders Molin wrote:
I think the project lead should pich his own structure
I think so too, but are there any basic roles you think need to be filled? Examples: Source Manager, Tech. Writer, QA Manager, etc.
Jason Henderson My articles
"The best argument against democracy is a five-minute conversation with the average voter." - Winston Churchill
|
|
|
|
|
Of course there is, but I don't know all of them before I see the tools we have to work with here on CP, and before I see the size/type of my team 
I hate when people make rules, saying we are going to do it this way, and then they force the team to work like that.
I love flexibility, and like to make the rules so they fit the team, this way your team is more happy, and if you do it right, you get a better product.
- Anders
Money talks, but all mine ever says is "Goodbye!"
|
|
|
|
|
Don't get me wrong, I don't want to force rules on you. These posts are just requests for comments that will help these projects get off the ground and on the right path.
Jason Henderson latest CPP news
"If you are going through hell, keep going." - Winston Churchill
|
|
|
|
|
Anders Molin wrote:
I'm not sure I really like that structure, it needs a lot of people.
Please, don't think of it as "structure". That was a bad term. Think of it as an offensive lineup, or the types of players that need to take the field during a pariticular play. I bascially gave an overview of an entire (american) football team. You need an offense, defense, and kick-off/punt/fieldgoal speciality teams, each customized for their task. The flexibility is there for project leads to do as they choose, but somehow in some form those roles need to be filled. I am not saying one person per role, several people will fill several roles, that is certain. I do however feel that you cannot deliver an SEI level 3 or higher product with any of those roles missing. (especially as the projects get larger and more ambitious)
Anders Molin wrote:
To be honest it looks like something they use at IBM or other big old compagnies
Relative to other areas of engineering disipline, software engineering is quite new, and is in need of some form of measurement of quality IMO. It's not intended as an imposing paradigm but as a standard against which to judge.
- Nitron
"Those that say a task is impossible shouldn't interrupt the ones who are doing it." - Chinese Proverb
|
|
|
|
|
OK, its been done before, but as I am a possible runner for PL in this project (subject to Tom Welch coming back), I wanted to get the ball rolling on this.
Any suggestions on the library and what should be in it can be posted on my Message board[^] as well.
So get going people. I saw an earlier post with a long list of suggestions, so I will be looking at that again as a good starting footprint.
Roger Allen
Sonork 100.10016
Were you different as a kid? Did you ever say "Ooohhh, shiny red" even once? - Paul Watson 11-February-2003
|
|
|
|
|
You're not really gonna call it UGLY are you?
I think GUILD or FLUID are better names.
------- signature starts
"...the staggering layers of obscenity in your statement make it a work of art on so many levels." - Jason Jystad, 10/26/2001
Please review the Legal Disclaimer in my bio.
------- signature ends
|
|
|
|
|
John Simmons / outlaw programmer wrote:
You're not really gonna call it UGLY are you?
I think GUILD or FLUID are better names.
UGLY is just a recognisable name for the moment, until a final formallised name gets decided on, its a good enough handle to use for the moment, which people recognise at that.
But after saying that, I will not do the project unless we can get the most annoying anacronym as possible!
Roger Allen
Sonork 100.10016
Were you different as a kid? Did you ever say "Ooohhh, shiny red" even once? - Paul Watson 11-February-2003
|
|
|
|
|
GUILD or FLUID are possible if there are no copyright issues.
How about this:
RUID - Roger's UserInterfaces Deluxe
I am sure it would boost Roger's motivation through the roof...
jhaga
CodeProject House, Paul Watson wrote:
...and the roar of John Simmons own personal Nascar in the garage. Meg flitting about taking photos.Chris having an heated arguement with Colin Davies and .S.Rod. over egian values. Nish manically typing *censur*. Duncan racing around after his pet *c.* Michael Martin and Bryce loudly yelling *c.* C.G. having a fit as Roger Wright loads up *c.* . Anna waving her *c.* and Deb scoffing chocolates in the corner.
...Good heavens!
|
|
|
|
|
For me it is most important that they are well tested om all relevant operation systems. I am prepared to do testing on XP.
[sorry for this anonymous]
jhaga
|
|
|
|
|
Anonymous wrote:
most important that they are well tested on all relevant operation systems.
This is a good point, we need to setup testers on various OS's. I myself have access to XP, 2K, NT4
Anonymous wrote:
[sorry for this anonymous]
As long as I know who to blame....
Roger Allen
Sonork 100.10016
Were you different as a kid? Did you ever say "Ooohhh, shiny red" even once? - Paul Watson 11-February-2003
|
|
|
|
|
I can test it on Win95, Win98 and Win2000.
-Dominik
_outp(0x64, 0xAD);
and
__asm mov al, 0xAD __asm out 0x64, al
do the same... but what do they do??
|
|
|
|
|
Please post all responses to Roger's message board as he asked.
thanks
Jason Henderson My articles
"The best argument against democracy is a five-minute conversation with the average voter." - Winston Churchill
|
|
|
|
|
Sorry, I have been away for a few days. In the near future my schedule will clear up but I'd hate to commit to anything that would let down a team project. I plan on being involved as much as I can, but I would certainly prefer that Roger take the PL on this project. I am always available if he is no longer available.
I am just amazed at the number of messages I need to catch up on!  Looks like I can spend all day on CP getting caught up.
Thanks for taking initiative on this, Roger.
|
|
|
|
|
check out my week 1 digest (top post at the moment)
Jason Henderson latest CPP news
"If you are going through hell, keep going." - Winston Churchill
|
|
|
|
|
Thanks for doing the digest!
"Blessed are the peacemakers, for they shall be called sons of God." - Jesus
"You must be the change you wish to see in the world." - Mahatma Gandhi
|
|
|
|
|
Looks like your the guy now.
Congrats!
Jason Henderson latest CPP news
"If you are going through hell, keep going." - Winston Churchill
|
|
|
|
|
I'm shamesly reusing Marc's idea
If any of you have some great ideas about the product(s) and how to make them, please make a post at my message board [^].
The original idea was to bild both a Defect Tracker (bug tracker) and a Soruce Control (like Source Safe or CVS) and make them work together.
My hope is to make a system that is as good as Bugzilla and as CVS, but make it work like Windows developers are used to. This means smoothless integration with VS.NET, and a good GUI for management and daily use.
Headlines:
What would you want in a new Defect Tracker
What would you want to see in a new Source Control System
What languages do we use to make this product
As I see it there's several parts in this project, and we might be able to use both native C++, and .NET and ASP.NET
- Anders
Money talks, but all mine ever says is "Goodbye!"
|
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
