Why Ford software sucks and why it will continue to suck

[DFS]

Translation:

FoMoCo contracted with the lowest bidders.
Musk engineers & builds everything in-house.
The results speak for themselves.

Well Musk imports a bunch of H1B “engineers” to artificially lower costs…

 

[EricM]

Well Musk imports a bunch of H1B “engineers” to artificially lower costs…

The quotes are legit. They are not engineers. They are coders.

AI will replace them.

AI is like a billion monkeys with a billion typewriters, eventually it'll spit out some good stuff that will be useable, and it can cull the crappy results itself.

A real software enigneer will have to validate the end product though.

 

[Muchmore]

A real software enigneer will have to validate the end product though.

PUH-LEEZ! That is what the the public is for----testing prod!!

 

[EricM]

Funny, but no. Ford simply can't put a steering rack out with shoddy software. People will die when it puts them over the hill, there will be lawsuits, and they will lose. Lawsuits are pretty much the only thing that can bankrupt Ford at this point. They have to produce code that is as good as what they had before at least- it cannot be junk or they will be liable when people die.

 

[smurfslayer]

The video & article don’t address the root issue. 7 development teams from several different companies isn’t uncommon. 7 independent “silos” of support is also common; but wrong. What’s missing is project management and that needs to happen with contractors and internal employees alike. In some ways, when you bring it in house you introduce new problems. There is always one team determined to make their lives simple and everyone else’s more difficult as a result. This leader needs to get whacked right up front, when they make the first move that affects all the other teams to bring them all into alignment. If you don’t, this team makes work life miserable for everyone else for YEARS to come, until someone audits the teams and identifies the problem children.

When contractors do this, you can fire them and someone else will step in and do it.

So bringing it in house is generally better, a bit more expensive up front, but less expensive in the long run, and also required diligent oversight. I hope they succeed.

 

[goblues38]

Disagree. Software is only as good as the requirements. Ford owns the product, therefore they control the requirements. The outsourced companies bidding for the work will build what they are asked to build. SO in this case, any issues are not the fault of the companies doing the work, it is the fault of the Product owner (FORD) telling them what to build.

I work in health care building software. Software Development Life Cycle (SDLC) is the same everywhere.

#1 - I want a widget
#2 - define the widget
#3 - develop the widget
#4 - test the widget
#5- deploy the widget

Very simplified example, but step #2 is what makes or breaks the success of the widget.

 

[thatJeepguy]

Well Musk imports a bunch of H1B “engineers” to artificially lower costs…

You spelled gas station owner/ operator wrong

 

[smurfslayer]

I won’t disagree, but will point out that a failure in step 2 makes it somewhat unlikely to pass step 4. Not impossible, bugs happen all the time. But, if it makes it past steps 2 & 3, then gets to test, it’s likely to fail because the test will be the manufacturer (Ford) and the work was done by Acme. Ford won’t tell Acme what the test is. If they did, that would mean that they like Acme and no employer likes contractors. Also, your SDLC is idealistic.

#1 - I want a widget
#2 - define the widget
#3 - develop the widget
#4 - test the widget
#4a - what do you mean the widget has to also do cartwheels?
#4b - redefine the widget
#4c - re-develop the widget
#4d - no, in fact I cannot provide you daily progress updates
#4e - ok, widget 1.1 is ready
#4f - test the wid^^ what do you mean we don’t have time for a full QA?
#5 - deploy the widget 1.1
#6 - onboard widget support cases
#7 - escalate widget issues to the widget engineering team
#8 - engage development to correct (some) widget flaws
#9 - patch the widget v1.2
#10 - deploy widget v1.2
repeat steps 6-10 until retirement.


Agree, Ford tells the contractors ‘hey, I want this widget, here’s what the widget needs to do. How much?’ Acme says ‘Great, we specialize in Widgets, but we only produce a few hundred widgets per year and you will have to place an order. Then we will need to see if you get an allocation. You may or may not get an allocation, you can check with our affiliate to determine the status of your widget allocation’. Turnaround time for Acme Widgets is 2 weeks to 104 weeks, tops. MSRP on the widget is only $9,999.99, which we are offering to Ford at 19,999.99, which is a SMOKING deal!’ ( I couldn’t resist ).

 

[smurfslayer]

You spelled gas station Stop n Rob owner/ operator wrong

FIFY ;-)

 

[GordoJay]

Disagree. Software is only as good as the requirements. Ford owns the product, therefore they control the requirements. The outsourced companies bidding for the work will build what they are asked to build. SO in this case, any issues are not the fault of the companies doing the work, it is the fault of the Product owner (FORD) telling them what to build.

I work in health care building software. Software Development Life Cycle (SDLC) is the same everywhere.

#1 - I want a widget
#2 - define the widget
#3 - develop the widget
#4 - test the widget
#5- deploy the widget

Very simplified example, but step #2 is what makes or breaks the success of the widget.

Defining the problem is what good engineers do well. Sadly, good engineers mostly want to go to Silicon Valley and get rich instead of going to Detroit and working for a salary at a big woke corporation that hates white guys. Sadly again, the pool of qualified non-white-guys is very limited and the bidding is intense. Ford can't compete for the cream of that crop. Apple and Google and Microsoft snap them up. Ford could improve things a bunch by chucking DEI out the window because there are competent unemployed white guys begging for jobs. Which do you think Ford will do? Scrape the bottom of the barrel or hire white guys? I'm hoping for the latter, but I'm not optimistic.