Re: Julia for Next-Generation Airborne Collision Avoidance System

[Optikos <ZUERCHER_Andreas@outlook.com>]

On Thursday, April 9, 2020 at 10:23:56 AM UTC-5, AdaMagica wrote:
> Am Donnerstag, 9. April 2020 07:44:55 UTC+2 schrieb J-P. Rosen:
> > I'd even say that the trend is for ease of writing rather than ease of
> > reading/maintaining. Well, software engineering
> 
> Would you really call this SE?

No, it isn't software engineering of large systems.  But mere talk of evangelism or of brow-beating shaming like what we in the Ada community have been doing for decades isn't winning the war and hasn't been winning very many major battles in recent years.

Instead, what could be done as an entirely different more-aggressive strategy to expose nonAda languages' problems more directly?  I am thinking something along the lines of what white-hat security researchers do to expose security vulnerabilities:  Ada white-hats exposing the vulns of nonAda software, contrasting exposed Julia weak spots in isolation with strength in Ada or Ada+SPARK replacements in isolation.

> > is not the only domain
> > where advertising for long term benefit against immediate gain is
> > difficult and unpopular...
> 
> “When Roman engineers built a bridge, they had to stand under it while the first legion marched across.
> If programmers today worked under similar ground rules, they might well find themselves getting much
> more interested in Ada!”
> Robert Dewar

There is really flying in the physical airplane.  And then there is simulating midair collisions of airplanes at fully-stressed-system scale in perfect-storm misbehaviors of Julia-based air-traffic-control software.  (Think what perfect-storm combination of situations must have occurred to trigger the Toyota crashes several years ago; use that analysis-technique thinking against Julia.)

Electrical engineering sets up labs with test jigs that stress the digital hardware in attempts at perfect-storm combinations of situations that must all be handled well without, say, excessive inductance, crosstalk, noise, heat, EMI, and so forth in the analog domain.  Ada community needs to set up a Julia test jig and an Ada test jig to compare them in controlled quite-harsh situations.