Discussion: April 1st, 2021 Development Update

What can I say… Welcome to the playground!

I would think some variety of hardware is of equal importance. Besides, even real world pilots seem to be able to disagree about various simming aspects.

3 Likes

I hope all you bug fixers will fix the old stuff in addition to the new stuff. After a week of trying, I got 500 Mb downloaded of my CRJ 550/700. I try a couple of times a day. One day I may reach 1.7 Gb and I can use what I paid for.

Well… I’d post the infamous quote Clint Eastwood used playing Dirty Harry about opinions and what they resemble, but it would get hidden for certain.

I think I’ll go fly my pretend airplane for a while in the pretend world of Flight Sim… or maybe read a book. :slight_smile:

I bought the Zinertek Airport Enhanced Airport Graphics yesterday (first marketplace purchase since it’s cheaper than on simmarket) and I find the differences quite nice.
The skid marks overlap the white runway marks (or the runway marks are a little transparent, which provides a little more realistic look on bigger runways.
The aprons look a little more cracked/used, it’s up to everyone’s taste. Gate positions have more oil/residue on them.

Works quite well with the CYVR from FSimStudios as well.

On some taxiways of default airports I noticed a repetitive pattern of the texture, I think some details added by this mod make it more obvious than the more generic textures from Asobo. But I don’t mind it since default textures are basically plain on many airports.

I think it was worth the buy, especially since it applies to basically all airports except for some payware, so it is an overall improvement even if it’s subtle.
Don’t expect groundbreaking results though. Also with some updates after it’s early release there is no noticeable fps impact for me, which is nice.

1 Like

In my opinion as former test director for multi million US dollars projects (you can believe me or not) it makes sense to split the community tester group in several sub groups along different dimensions IF you really what to have the maximum output - not getting 300 times the same bug report.

First dimension: Experience from newbie to heavy modder
Second dimension: Hardware from laptop with low power graphics card/CPU up to very expensive NVidia 3080
Third dimension: Airplane from small and slow (C152) up to airliners
Fourth dimension: Realismus setting from have fun to fly up to fly for hours with ATC, flight plan, real weather.
More dimensions …

If you split 300 people into groups, every group will have only 10 to 30 participants. And it will take some time to select them. But this is then a HETEROGENOUS tester group, something you NEED if you want to do real testing, and not only say “we listen to you (and now don’t bother us any longer)”.

3 Likes

Man, you speak the truth! But “agile” is now for some years the new “silver bullet”.
Do we need analysis and design?
Do we have to walk down the decision tree?
Do we have to verify our ideas with prototyping?

The sad answer to all these questions: No, we don’t, we have agile.

Not only Asobo/Microsoft have this illness, the company I work for has the illness, too.

Don’t get me wrong: I work agile for 20+ years (yes, the method was around before that famous meeting), but agile is only one (little) tool in my tool set.

1 Like

Don’t get frustrated and don’t feed the trolls. You know the drill.

This is the sad thing about automated testing: You have to write the software and you have to write the test for the software. But everywhere are bugs. The bug can be in your software, in your test software, in your model or somewhere else.
Automated testing is fine IF you have a baseline. But I say: MSFS 2020 has no baseline, yet. Or - for example - do you really want that the current turboprop behavior is taken as baseline?

1 Like

I think you expect wrong. Agile says (at least for managers): Release when the clock ticks. Bug fixes are in the next release (or in the release after the next release, or after, …)

Just remember: Some bugs like the forever #1 bug Garmin are in the list since August 2020 and were in the list before first release.

1 Like

Okay. I start immediatly on my resume. But stop, who will read my resume and decide? The same people that brought us the “flaps lift” bug or the “76T Bishop one frame per second” bug?

My advise: Put some sweets or “I love you” stickers into your resume to make decision easy for the decision makers.

1 Like

The dilemma of software engineering. You witnessed it here. First.

4 Likes

You have absolutely no idea about automated testing.

2 Likes

That is for sure. I try every day to download my new CRJ550/700. A couple of times I have downloaded close to 700 Mb before it starts from the beginning again. Looking for the last 1.1 Gb

Hi. It had to do with “very, very disappointed” and “publicity stunt”.
I was thinking: Asobo is showing that they are moving in a different direction and still it’s not enough. Still it’s disappointing and untrustworthy according to some people here.
I just happen to perceive this differently. I see it this way: Asobo has taken this journey on an untrodden path and we already can travel with them and enjoy and marvel at everything that already works fine. Yes, there are (many?) points to smooth out and from what I read here some serious annoyances. But there’s already so much very, very good and Asobo shows again and again that they take the community seriously. So I think that’s why I reacted to your post. All the best!

4 Likes

You should at least say what is wrong with his idea. Just stating he has no idea is not very constructive.

2 Likes

Well, there is so much wrong with this. For starters, integration tests are supposed to test one well defined aspect of the product. „Flying around for hours“ and „detect memory leaks“ is not well defined (in terms of computer science).

Second, integration tests need to be „practical“ in the sense that they tell you as fast as possible that something broke. „Hours“ (for a single „fly around“ test) is not practical: you simply do not want each developer to wait for hours until his/her change gets integrated.

Of course you could run those tests overnight. But then who is to blame? Which one out of the hundreds of checked in changes during the day is responsible? Again: not practical.

A well-designed integration test tests one thing (aspect/functionality/module/…), and if it breaks you immediately know what is wrong (that is not always true or practical in reality either, but is the gold standard every integration test should aim for).

I could go on for a while (and if you were a software engineer you‘d know what I was talking about), but you get the idea…

P.S. To clarify: I did not intend to „discredit“ the author to which I initially replied, for not knowing how automated testing works. My tone was too rude and for this I apologise.

I was simply carried away (once again) by the naiveté and ignorance - and sometimes outright angry comments which seem to be targeted at Asobo employees - about software engineering and expectations that people seem to have here in the forum (again: not the poster to which I replied!). And the same „negativity“ surfaced here again in this thread.

I fully understand the frustration that some people have, but get real, people: everyone who claims to have „special rights or justified expectations“ just because „they have been into flight simming for years“ needs to adjust their reality a little bit. You did not purchase a flight simulator for hundreds of thousands of dollars with dedicated support comtracts (which cost extra!): you purchased a „piece of entertainment“ - a „computer game“. With no obligations whatsoever from the vendor (except perhaps if the software would willingly brick your PC or whatever. But IANAL).

Microsoft and Asobo have commercial interests: so they might promote their Xbox offer more than investing time into fixing (finishing) some autopilot X in aircraft Y.

But beyond all that: software engineering is a complex matter! Yes, you can produce software in various degrees of quality (simply put), but even what might be very obvious for every user here in the forum (like the flaps issue) might simply be an unfortunate side-effect of a „last minute checkin“.

And to all those who cry out „show stopper! I can‘t play this game anymore!“: personally I am counting more improvements with each update than regressions. I am fine with that development. Period.

11 Likes

And that is a problem, too: you get a lot of redundant bug reports that need to be sorted and (re-)classified, and potentially also lots of false reports. (*)

Quantity does not necessarily improve quality, especially if you have limited resources to deal with it.

(*) A notable example here in the forums was that guy with that click-baity title like „London looks terrible!“: he went full berserk on how London looked nothing like he expected and he even claimed to have checked all settings.

After a couple of posts someone asked „Have you really enabled photogrammetry, by enabling Bing data downloads?“ And the guy‘s reply was: „What is photogrammetry? Where do I enable it?!“

Well, it turned out that he had that option disabled (probably due do the recent update, but that is another story). The point to note here is that a) he insisted to have checked all his settings and b) his brain somehow cancelled out screenshots that other people had posted how London was supposed to look (with enabled photogrammetry), but no, he insisted that London looked terrible and that hence the UK update was „worthless“.

So that is a problem if you open up a test program to too many people: you have to identify and sort out all those false rejects.

Don‘t get me wrong: if you think that something is wrong, then by all means report it (after having checked whether the same issue might have been reported). I am just saying that there is always a certain percentage of „bogus reports“, and the larger the absolute numbers the more time it takes to sort them out.

5 Likes

Totally agree.

I don’t understand why people think the invited beta testers need any qualifications other than to be enthusiastic about the simulator…

I don’t believe for a second that we are all expected to be computer programmers, game designers, real world pilots, etc. And those attributes might well prove to be counterproductive anyway.

I’m expecting them to ask us to “fly the sim”, ideally around the new world update area, and let them know if we encounter any new significant issues.

IMHO the beta is here to find major flaws such as the flaps bug and major performance drops.

Anything else, minor flight model changes, small POI defects, etc etc, would be found by the community after release anyway and be fixed later.

My experience or lack of as a real world pilot is besides the point.

4 Likes

Not just that: a stress test is what it is called: an attempt to find potential performance bottlenecks, mostly in your server infrastructure. And in such a stress test scenario you want to create load on your servers, created by as many „real world clients“ as possible (you can always synthesise load algorithmically, but that might statistically behave differently than „the real world“, and you might miss that critical performance bottleneck: nothing beats the creativity of a gamer that wants to bring down your servers :wink:

This FS test on the other hand is not a stress test. And as noted above a lot of bug reports do not necessarily help either (in a timeboxed test/release window).

2 Likes

Exactly this!

This beta is not about identifying some missing/misbehaving submenu in some autopilot.

It is about to identify - and hopefully prevent - the most obvious „showstoppers“.

And for this a good mix of users (from real world pilots to fun flyers) is the best. The only qualification should be to be motivated and be able to write concise bug reports.

3 Likes