DataStage delivery
Moderators: chulett, rschirm, roy
DataStage delivery
Which could deliver a complete & quality ETL solution with DataStage the fastest?
-
vmcburney
- Participant
- Posts: 3593
- Joined: Thu Jan 23, 2003 5:25 pm
- Location: Australia, Melbourne
- Contact:
Of course you've got to add a disclaimer that most of the respondents are experienced datastage users and could be baised. :D
From experience the mistakes made by new DataStage users, and I made a ton of them myself, cause ongoing productivity losses for developers, testers and support staff for the life of the project, including post implementation. So while a large number of inexperienced developers could create the ETL processes in the shortest amount of time these gains would be lost in other parts of the project.
These problems are significantly reduced if there is an experienced DataStage resource available to do quality assurance and mentoring for the inexperienced users. A combination of new and experienced is probably the easiest to assemble.
From experience the mistakes made by new DataStage users, and I made a ton of them myself, cause ongoing productivity losses for developers, testers and support staff for the life of the project, including post implementation. So while a large number of inexperienced developers could create the ETL processes in the shortest amount of time these gains would be lost in other parts of the project.
These problems are significantly reduced if there is an experienced DataStage resource available to do quality assurance and mentoring for the inexperienced users. A combination of new and experienced is probably the easiest to assemble.
Certus Solutions
Blog: Tooling Around in the InfoSphere
Twitter: @vmcburney
LinkedIn:Vincent McBurney LinkedIn
Blog: Tooling Around in the InfoSphere
Twitter: @vmcburney
LinkedIn:Vincent McBurney LinkedIn
I think vmcburney makes a good point in this:
"A combination of new and experienced is probably the easiest to assemble." I think a mentoring type arrangement can work well.
One observation... the new people should have some sort of client/server, distributed environment aptitude or the project may suffer. The centralized mentality (i.e. mainframe) is a tough paradigm to overcome for some people.
"A combination of new and experienced is probably the easiest to assemble." I think a mentoring type arrangement can work well.
One observation... the new people should have some sort of client/server, distributed environment aptitude or the project may suffer. The centralized mentality (i.e. mainframe) is a tough paradigm to overcome for some people.
Kevin Janes
-
ray.wurlod
- Participant
- Posts: 54607
- Joined: Wed Oct 23, 2002 10:52 pm
- Location: Sydney, Australia
- Contact:
The question really misses the point.
What is really vital is knowledge of the business goals of the process, preparation of quality metadata (which includes complete design of the ETL process) and adequate preparatory training of the developers.
There will probably be cases where the "fastest" solution would be ONE experienced developer. In other cases any of the possibilities might be appropriate, though I'd be wary of an all-inexperienced crew.
Accordingly I won't be voting.
What is really vital is knowledge of the business goals of the process, preparation of quality metadata (which includes complete design of the ETL process) and adequate preparatory training of the developers.
There will probably be cases where the "fastest" solution would be ONE experienced developer. In other cases any of the possibilities might be appropriate, though I'd be wary of an all-inexperienced crew.
Accordingly I won't be voting.
A few thoughts
1) I recently saw in a TDWI white paper an estimate that it takes 3 months to learn an ETL tool and 1 year to become good at it. Just an opinion, but I think it is fairly accurate.
2) I've also thought the best ETL team would consist of experienced and new ETL developers. In theory the new developers could be billed at a lower rate (if the ETL work is outsourced) saving the client money. Also, there are many basic tasks the experienced developers don't want to waste time on, but are good things for a learner to be doing.
3) I also think it is interesting that there is a threshold of value when adding new developers. There are times when adding a new developer or two in the middle of a project would make a little impact in when the work is completed, and can sometimes even delay it
2) I've also thought the best ETL team would consist of experienced and new ETL developers. In theory the new developers could be billed at a lower rate (if the ETL work is outsourced) saving the client money. Also, there are many basic tasks the experienced developers don't want to waste time on, but are good things for a learner to be doing.
3) I also think it is interesting that there is a threshold of value when adding new developers. There are times when adding a new developer or two in the middle of a project would make a little impact in when the work is completed, and can sometimes even delay it