Reply to: Another EDS success story

[DeepBlue]

From my experience of poublic sector organisations, its usually the management at the organisation that cause all the delays with counter-productive bickering, requirements change, indecisiveness and no clear leadership in the direction of the project. Every manager of every department sticks their oar in to criticise requirements defined by other departments and its a complete mess.

Sounds familiar! Mirror that sort of effectiveness within your own organisation and damn you've got a money spinner of a lame duck.

[JoJoGabor]

From my experience of poublic sector organisations, its usually the management at the organisation that cause all the delays with counter-productive bickering, requirements change, indecisiveness and no clear leadership in the direction of the project. Every manager of every department sticks their oar in to criticise requirements defined by other departments and its a complete mess.

I have witnessed Microsoft tell the main stakeholder of a project that they are asking for far too much design effort for a certain project, yet the public org insisted on using them for all sorts of design exercises at an extortionate cost as they insisted that everyone involved was a senior level consultant - even for the frunt work!!

[DiscoStu]

I thought you were in prison!

Oh, no sorry, that's the other bulltulip merchant.

[Churchill]

They could have bought one from my company.

It works too.

Which maybe is why they didn't...

I thought you were in prison!

Oh, no sorry, that's the other bulltulip merchant.

[threaded]

They could have bought one from my company.

It works too.

Which maybe is why they didn't...

[cojak]

Well... when people say "database" that's not really what they mean.

A database, yes, that could be done cheaply. Design a schema and implement it in a scalable and secure fashion, not too hard. However a database is only as good as the data in it...

For the data in the database to be useful we have a) data migration b) user updates and c) reporting.

Data migration is a serious pain in the rear end. The incoming data is never clean. Never ever. So if you want to enforce some validation etc. in your new database, you're going to spend a lot of time cleaning up the old data. Don't think this is a job that you can pay Indian Development Centre to do. Been there, seen the damage. Sadly explaining what an "intelligent decision" is to most of those guys takes about 6 weeks of "FAIL. RETRY".

Once the data is there, we come to the meat of the project. Processes. Processes processes processes. Everyone has their own "pain point" which they hate, and wrote into the proposal to be fixed. Most of the processes will be straight forwards. Transfer of prisoner Gordo Brown from no. 10 to Holloway MS. Transfer of prisoner to solitary confinement. etc. etc. But someone will want workflow, so that the transfer is requested, approved, actioned. (All very sensible).

Now if you start to think of all the possible state transitions of a prisoner/parolee in the system, and assume each transition will have at least one request screen, one approval screen, one actioned screen... you see where the work might end up.

Then we look at reporting. Everyone with "manager" in their title will want their own report. Kerching for EDS. "Oh yes, we can make individual reports, really fast and easy. We can train your managers(*) to allow them to create their own reports."

(* Assuming said manager has programming language/database experience...).

An army of contractors making stupid reports that add no value to the system at all. Guestimate of 400 reports generated for such a system, and only 40 of which are actually used to add value.


So yeah. Not just a DB of Record. That's easy. Customers assume processes around their data, and they are expensive.

Post of the day there Deepblue.

Take note fellow newbies - that's the way to make an entrance^^^^^^^

[DeepBlue]

If you were desperate enougn to accept such a prohect as above - make sure youve got indemnitry as it could very well end in tears - and guess who would be left holding the poison chalice ?

Thanks Alfred. It was pretty close to the top of my list when looking at this whole contracting endeavour. Have driven cars without being insured before (oops, but it was an emergency, um... yeah). Didn't like that one bit as I was fully aware I was putting my entire wealth in my hands :-S

[xoggoth]

Indeed, db thanks for the explanation. Actually, a lot of that sounds pretty familiar from projects I have worked on for the MOD.

[AlfredJPruffock]

Most of my experience has been in data oriented integration. It's never as simple as it looks. Thankfully I've steered away from financial systems so far. Have been involved with one integration point in a project, where the customer didn't really know which financial system was correct (asset allocation between divisions). And scarily, they didn't appear to care that finance system in accounts had data that looked nothing like the ERP data "well users from A will complain it's wrong, and users from B will complain its wrong... can you fix it?"

If you were desperate enougn to accept such a prohect as above - make sure youve got indemnitry as it could very well end in tears - and guess who would be left holding the poison chalice ?

[DeepBlue]

I'm just coming to the end of an 8 month project installing an MRP system and the biggest task, by far, has been validating data from the current financial system that its linked to. Holiday from January

Most of my experience has been in data oriented integration. It's never as simple as it looks. Thankfully I've steered away from financial systems so far. Have been involved with one integration point in a project, where the customer didn't really know which financial system was correct (asset allocation between divisions). And scarily, they didn't appear to care that finance system in accounts had data that looked nothing like the ERP data "well users from A will complain it's wrong, and users from B will complain its wrong... can you fix it?"

[Platypus]

by the time the poor guy on the end gets his renumeration

I am not a number !

(perhaps you meant remuneration ?)

[Sysman]

I knew (in the biblical sense) an EDS 'consultant' for a couple of years, on and off. Incompetent angry nutjob, and very defensive of EDS! Very proud of her EDS keyfob she was.

Not an ex Computer Associates type by any chance?

[TinTrump]

Good reply DeepBlue. My main area is MRP/ERP systems. One thing I try to get a business to do is agree a standard set of reports required from a system. That way they get a consistent measure of how they're doing in terms of product delivery and they don't have every dept. doing their own thing. Simpler too.

I'm just coming to the end of an 8 month project installing an MRP system and the biggest task, by far, has been validating data from the current financial system that its linked to. Holiday from January

[DeepBlue]

The only database work I have done was a purely technical one in Access, which people on here will be quick to point out is not a database at all.

But in what way is one of these huge database systems substantially different from another? A lot of linked PCs in police stations/doctor's surgeries/job centres etc. holding names and addresses of arrestees/patients/job seekers etc. with details of alleged offences/treatments/applications etc.

How can projects which must have so many simularities keep failing over and over? Or is there some complexity us non experts don't know about? Can any database experts explain it?

Well... when people say "database" that's not really what they mean.

A database, yes, that could be done cheaply. Design a schema and implement it in a scalable and secure fashion, not too hard. However a database is only as good as the data in it...

For the data in the database to be useful we have a) data migration b) user updates and c) reporting.

Data migration is a serious pain in the rear end. The incoming data is never clean. Never ever. So if you want to enforce some validation etc. in your new database, you're going to spend a lot of time cleaning up the old data. Don't think this is a job that you can pay Indian Development Centre to do. Been there, seen the damage. Sadly explaining what an "intelligent decision" is to most of those guys takes about 6 weeks of "FAIL. RETRY".

Once the data is there, we come to the meat of the project. Processes. Processes processes processes. Everyone has their own "pain point" which they hate, and wrote into the proposal to be fixed. Most of the processes will be straight forwards. Transfer of prisoner Gordo Brown from no. 10 to Holloway MS. Transfer of prisoner to solitary confinement. etc. etc. But someone will want workflow, so that the transfer is requested, approved, actioned. (All very sensible).

Now if you start to think of all the possible state transitions of a prisoner/parolee in the system, and assume each transition will have at least one request screen, one approval screen, one actioned screen... you see where the work might end up.

Then we look at reporting. Everyone with "manager" in their title will want their own report. Kerching for EDS. "Oh yes, we can make individual reports, really fast and easy. We can train your managers(*) to allow them to create their own reports."

(* Assuming said manager has programming language/database experience...).

An army of contractors making stupid reports that add no value to the system at all. Guestimate of 400 reports generated for such a system, and only 40 of which are actually used to add value.


So yeah. Not just a DB of Record. That's easy. Customers assume processes around their data, and they are expensive.

[DeepBlue]

Always makes me happy that I avoided a consulting job with one of these types of companies. Going the solo contractor/consulting route after putting in 4yr+ with a small somewhat competent software co.

EDS? Worst of a bad bunch.