Preach on.

On 16 Sep. 2016 10:50 am, "David Connors" <[email protected]> wrote:

> On Fri, 16 Sep 2016 at 10:33 Greg Keogh <[email protected]> wrote:
>
>> What do you suggest as an alternative?
>>>
>>> Writing stored procedures.
>>>
>>
>> What about the classical problem of "impedance mismatch". You have to
>> carefully maintain DataSets or similar and use DataAdapter to fill them,
>> then writing data back is a circus trick with the ADO.NET classes.
>>
>
> Otherwise referred to as doing high quality work.
>
>
>> Then they invented ORMs, why did they do that!? -- *GK*
>>
>
> I have given this considerable thought over the years. Normally I explain
> this with swear words but I think it boils down to two key factors.
>
> *Weltanschauung*
> The people who think that ORMs are a good idea have a code-centric view of
> the world. Careful declarative design of a data tier is outside of this
> world-view so they see it as overhead (plus they often have to bargain with
> smelly neck beard DBAs).
>
> *Free Lunch / Laziness / Lack of care for end result*
> Developers get excited over the prospect of auto-generation because OMFG
> look at all that code I did not write actually. Most developers don't wear
> the ops cost of their solutions and they certainly don't USE them and
> consequently don't give a toss if some EF-based turd they engineered takes
> 10 seconds to do things that should take 10msec.
>
> Things would be different if the average engineer were forced to walk a
> mile in the average IT pro / end users shoes.
>
> YMMV but I had a good rant on this a few years back:
> https://www.youtube.com/watch?v=YMfRahO8fLo (jesus that was 6 years ago).
> I have softened my stance on Agile somewhat since then.
>
> Good outcomes always take more effort and energy. The universe has been
> that way for 13.8 billion years and isn't going to change any time soon.
>
> David.
>
> --
> David Connors
> [email protected] | @davidconnors | LinkedIn | +61 417 189 363
>

Reply via email to