Reply to: Our highly educated grads on the scrapheap

[Mich the Tester]

Quite so. When I was writing real code, up to and including single-handed implementation of accounting sub-systems, I was well aware that I was building a wheel from scratch: many others would have done exactly the same thing before, usually the same way.

I didn't really have a problem with that: I've long regarded software writing as a kind of engineering, and an engineer doesn't expect to be the first to face the same problems (but still nobody calls his work cut'n'paste because it isn't).

Now it is indeed true that the common elements are more systematised, and I personally make a living configuring rather than coding. ISTM that it still takes many of the same skills, with the exception that you can configure badly and get away with it. But it still needs specialists.

Yes. As a tester I come across apps that have been built by a tool monkey and apps that have been built from scratch by an experienced programmer who migt use a tool just to help him here and there but who will go through his code checking for faults and cleaning up the spaghetti that the tool produces. The first category are no fun at all; you find bugs too easily and there's no challenge to it. Usually you'll find that the tool monkey understands nothing about internal memory and when the tester forces an error message using incorrect input he can screw up the whole app, because the input can't be corrected properly; the tool monkey doesn't understand the need to clear up old instances. The second category is far more challenging and actually makes testing enjoyable; trouble is you can go for days without reporting errors and then some numpty 'managerman' with no IT training starts moaning that 'there are not enough bugs being found'.

[expat]

I think that part of the problem is the increasing generic nature of business processes and information systems.
..... Its all been done, reducing so-called software "engineers" to software "implementors". Anybody who has the time or inclination to design their own framework will invariably run into 90% of the design and implementation issues that the other opensource framework ran into.

In a nutshell, 50% of what was a software design and development task 8 years ago is now a configuration and implementation task.....

Quite so. When I was writing real code, up to and including single-handed implementation of accounting sub-systems, I was well aware that I was building a wheel from scratch: many others would have done exactly the same thing before, usually the same way.

I didn't really have a problem with that: I've long regarded software writing as a kind of engineering, and an engineer doesn't expect to be the first to face the same problems (but still nobody calls his work cut'n'paste because it isn't).

Now it is indeed true that the common elements are more systematised, and I personally make a living configuring rather than coding. ISTM that it still takes many of the same skills, with the exception that you can configure badly and get away with it. But it still needs specialists.

[Menelaus]

When you put it that way we really should moan a little bit less.

Not sure, tbh. All generalisations are dangerous, including this one, remember?

[Mich the Tester]

I'm not sure that this subject should develop into an intellectual pissing contest.

By mere dint of our being predominantly (generalisation alert!):

- science graduates
- self-employed
- upper 10% earnings

...we're already near the top of our individual communities, irrespective of how we choose to regard them.

When you put it that way we really should moan a little bit less.

[Pondlife]

Oh no, said the agent unselfconsciously, after 6 months experience I'd put someone in as a senior consultant.

Oh Dear (tm)

I suppose if you can update your contacts in Outlook, Siebel's only a conversion course away.

[expat]

Oi! Come over here and say that!

Seriously though, testing's got a bad name thanks to insultancy companies sending humanities graduates on two week courses and then selling them to clients as 'test consultants'.

Nobody who has worked with good testers would think them a waste of space. But ISTM that the muppets who run this trade can't fit that into their model, hence the feeling that anybody could do it.

Not that it only applies to testers. One agency that I talked to once questioned my self-rating of my ability in Siebel: with almost 2 years experience, in only 1 environment, I blushed at calling myself a consultant. Oh no, said the agent unselfconsciously, after 6 months experience I'd put someone in as a senior consultant.

[Menelaus]

I'm not sure that this subject should develop into an intellectual pissing contest.

By mere dint of our being predominantly (generalisation alert!):

- science graduates
- self-employed
- upper 10% earnings

...we're already near the top of our individual communities, irrespective of how we choose to regard them.

[sunnysan]

IT and automation

You guys are all delusional aren't you? Since when has IT required a lot of skill? It's precisely because the barriers to entry are so low that you guys are suffering now.

I think that part of the problem is the increasing generic nature of business processes and information systems.

As in any maturing industry, patterns are becoming apparent and these patterns are implemented by software vendors as frameworks(persistence, GUI, Security ) or as applications, and the biggest impact is SAAS related technologies such as SalesForce.

If we take SalesForce as a case in point, in order to utilise the technology to meet the needs of the organisation the skills required are understanding the business and not the technology.

To a lesser degree, open source frameworks perform the same function at a lower level. Why design a ORM,security,WEB front end framework these days? Its all been done, reducing so-called software "engineers" to software "implementors". Anybody who has the time or inclination to design their own framework will invariably run into 90% of the design and implementation issues that the other opensource framework ran into.

In a nutshell, 50% of what was a software design and development task 8 years ago is now a configuration and implementation task.

There will always be a need for top flight software engineers, but the need is decreasing in the general marketplace and the value of the skills decreasing.

What does the future hold? I dont know, but I do beleive that the quest for management metrics and the ability to deliver and measure with maturing technologies are going to reduce the white collar workplace into a dull automated environment , with employees being tied to their email and BPM systems which can be set up and configured by people with little or no technical skill.

I am not sure of the details, but the so-called "industry of the future, IT" which was thought to be a solid career choice 10 years ago, will be devoured by the same forces that spawned it.

[Mich the Tester]

That's pretty much what I was itching to type in reply.

The idea of some jobs needing highly-skilled people does not fit with the agency model (or the HR model for all I know), where:

• a "skill" is represented by a single keyword
• a job requires a "skill", or would prefer to have it, or doesn't need it
• a person "has" a "skill", or doesn't.

This allows dummies to match "skills" (where "skill" = keyword), but doesn't allow for work being done by the skilled people who should be doing it.

I call this Tool Oriented HR, and it's directly related to Tool Oriented Testing and Tool Oriented Development.

A fool with a tool is still a fool.

[expat]

I beg to differ. It depends what you do, in IT.

There is also two distinct tiers of IT work out there. Good stuff that attracts top tier people who could otherwise be lawyers, doctors (20%) and the rest (80%) that attracts those that would otherwise do non-professional roles if it weren't for IT.
.....
My point is, that the top tier IT stuff is highly skilled work done by professionals, its the 80% that drags down the entry level and perception of the job.

That is a fair point but I would disagree with the conclusions you draw.

IT is a skilled business. Those of us who have been doing it for a while know about the complexities involved with all stages of the lifecycle and all roles involved in any project.

The problem is (as I see it) that the rest of the world seems to think that anybody can do it and will employ anybody if they are cheap enough.
.....

That's pretty much what I was itching to type in reply.

The idea of some jobs needing highly-skilled people does not fit with the agency model (or the HR model for all I know), where:

• a "skill" is represented by a single keyword
• a job requires a "skill", or would prefer to have it, or doesn't need it
• a person "has" a "skill", or doesn't.

This allows dummies to match "skills" (where "skill" = keyword), but doesn't allow for work being done by the skilled people who should be doing it.

[Mich the Tester]

Most of the Java devs I know are pretty good, and they are bright people. Some though are crap (including some of my current colleagues!) But overall developers do actually have to be capable of developing something. It's all the other hanger-onners who drag down IT. I'm talking testers, sysadmins, infrastructure people, BAs, TAs, PMs, 'architects'. Whilst you can find good ones of these, many are of questionable value or actually no value.

Oi! Come over here and say that!

Seriously though, testing's got a bad name thanks to insultancy companies sending humanities graduates on two week courses and then selling them to clients as 'test consultants'.

I may be blowing my own trumpet by saying that to be a decent tester you need to have been a developer first. You don't necessarily need to have been a good developer, but need the experience of developing from an often crappy spec so that you know what kind of errors developers and those pesky 'business analysts' make, and just as importantly, have the technical skills to analyse a fault and thereby help the developers in solving it.

Unfortunately though I feel that the way many IT departments and projects are run, unskilled, unthinking automaton testers are actually at a commercial advantage; just get the right certificate and do your work according to clientco's 'best practises' and gee whizz, you can call yourself a tester and get yourself a juicy salary.

[SillyMilly]

It's all the other hanger-onners who drag down IT. I'm talking testers, sysadmins, infrastructure people, BAs, TAs, PMs, 'architects'. Whilst you can find good ones of these, many are of questionable value or actually no value.

Clearly your coding errors have been exposed by testers a bit too often!!

[The Lone Gunman]

It's precisely because the barriers to entry are so low that you guys are suffering now.

That is a fair point but I would disagree with the conclusions you draw.

IT is a skilled business. Those of us who have been doing it for a while know about the complexities involved with all stages of the lifecycle and all roles involved in any project.

The problem is (as I see it) that the rest of the world seems to think that anybody can do it and will employ anybody if they are cheap enough.
There are no barriers to entry into the business.
How many other businesses would allow unproven, unqualified hobbyists to be core to (often) multi million pound projects?

I am a reasonably good home mechanic and have built a few club rally cars. Do you think I could get Ross Brawns job?
That may not be a sound example but I have worked on major projects where much of the team is there because they have tinkered a bit at home.

Also, the term IT covers a multitude of sins. What I am talking about comes under the Software Engineering banner.
IT is an engineering discipline and should require practitioners to be properly qualified and/or experienced

[aussielong]

Most of the Java devs I know are pretty good, and they are bright people. Some though are crap (including some of my current colleagues!) But overall developers do actually have to be capable of developing something. It's all the other hanger-onners who drag down IT. I'm talking testers, sysadmins, infrastructure people, BAs, TAs, PMs, 'architects'. Whilst you can find good ones of these, many are of questionable value or actually no value.

BA - a non-job. Get rid of the job and employ Developers that can talk to the business. ie. "Analyst Programmer" 's. BA takes half the fun out of programming. But they often don't do a good enough job to help me. I need to dig in and understand everything from the source to be able to implement their spec - so many holes. Mainly failed developers.

I have only met one great BA. He was soooo useful. It made programming somewhat boring but we got a lot more work done in the team because of him.

[swamp]

I beg to differ. It depends what you do, in IT.

There is also two distinct tiers of IT work out there. Good stuff that attracts top tier people who could otherwise be lawyers, doctors (20%) and the rest (80%) that attracts those that would otherwise do non-professional roles if it weren't for IT.

C++ development does not have a low barrier to entry - and does indeed require a high degree of skill - otherwise your stuff just will not work (for long) and you will be exposed quickly. Natural selection keeps the standard high and prevents outsourcing to armies of new grads overseas. That said, its dying because of hardware advancements.

Java on the other hand at the low end is very easy to enter and this drags down the average. The pool of Java jobs is bigger than C++ but within that pool only 20-30% of the work would attract top flight devs.

Only the top 20% of IT jobs are worth doing from the point of view of a career. 80% is dull blue collar work without prospects, stressful and not worth the pay.

I'm talking from Dev POV since that's what I do. Having dabbled in Sys Admin, Network Admin for a year back at the start of my career - I would say you find even more cowboys doing those roles than Dev.

My point is, that the top tier IT stuff is highly skilled work done by professionals, its the 80% that drags down the entry level and perception of the job.

Most of the Java devs I know are pretty good, and they are bright people. Some though are crap (including some of my current colleagues!) But overall developers do actually have to be capable of developing something. It's all the other hanger-onners who drag down IT. I'm talking testers, sysadmins, infrastructure people, BAs, TAs, PMs, 'architects'. Whilst you can find good ones of these, many are of questionable value or actually no value.