bfr

Did it actually genuinely stop working beyond that point or is that just the release when ED said they'd no longer support any issues that might arise if people were on that OS?

So a microkernel and not really a fully fledged platform/OS, and one for which nothing much really exists outside of the embedded space. Right-o. Anyway, shall leave it at that given we've meandered quite a bit off topic here.

Out of interest, can you name for me a platform or platforms that you consider to meet your bulletproof standard?

Really? I can't say i've ever pined for the old one back. It was briefly less familiar but i'm struggling to think of the old one did that this one doesn't. You'll have to ask ED what their motivation to change was, perhaps support was a consideration. Yes, upgrades often aren't risk free and are usually a risks v benefits thing. - Am I going to get better features/performance/outcomes by doing this? - Is it going to be easier to support than what I currently have? If both the answers are no then don't bother by all means as there is genuinely no point. Usually though one or both things are true. I work on an in-house team for a company. We don't directly sell what we write but the business very much depends on it to provide a service and make money so we all get paid. Sustainability is hugely important and the bit you seem to miss is that getting rid of old stuff that is difficult to maintain and support going forwards is often an integral part of sustaining that. We innovate because our competition does and thus if we don't then we ultimately lose out as a business. Weird flex. Software can and often does take ages to write and its not like every single traditional engineering project is revolutionary/designed completely from scratch rather than iterative. It has always been with great effort and careful management. Its entirely possible to balls up a Unix system or whatever other platform you care to mention by haphazardly setting things up or not keeping on top of things like patches and updates. God knows i've seen enough of them.

As i'd mentioned in earlier posts, i've had both the Harrier and Strike Eagle since launch and I would've finally bought the Mirage 2000 but for the beginnings of disquiet between ED and RB meaning their older modules never appeared in a sale again. So yes, I do have skin in the game as and when these modules cease to work.

Good luck getting replacement parts in a hurry for that IBM XT that you've had chugging away running MS DOS 3 in the corner since 1990 and now means a critical link in a business process is dead in the water until its sorted. Every chain is only as strong as its weakest link and unsupported legacy hardware and software can be a very weak one indeed. I can drive around in a 50 year old car if I wanted and it'll get me from A to B whilst it still works. I have to accept though that if the car does break down then parts might be difficult to impossible to source and I might be off the road for a while and it might cost me a small fortune to get back on it. Bad design is and has always been and always will be a thing. You're right that newer isn't automatically better but there are countless things that have improved in terms of user experience by successful iterations. And I can similarly look back at things i've worked on and with in the past and think 'what the hell were people thinking back then?'. Those are all things that can be done on modern systems provided you know how to set them up and manage them properly. People are always part of the overall system and (again making the chain analogy) usually the weak link e.g. ransomware doesn't just appear but rather someone opened the door to it. If its then able to get a foothold and cause havoc through being able to run at an extremely elevated level of privilege then that can quite possibly be down to someone being given privileges beyond what they should've been running under. That isn't remotely a new phenomenon. Do you think software engineers 'move fast and break things' because they actively desire to or because people who pay theirs wages often demand the moon on a stick by Wednesday morning?

Funnily enough you're talking to an ex-AS/400 dev and admin. Try getting devs for that platform if you've got a big project coming up. You'll find them surprisingly few in number and surprisingly expensive because its now incredibly niche. I moved away to other stuff years ago because the market was only going one way. Now try getting support for that 30 year old ERP suite you're still running. Or getting parts for your long since out of production mainframe when something goes pop. An example would be when I first started out in my IT career my then employer was still using the AS/400's predecessor (System/38). We got rid of that for an AS/400 because: - The OS and ERP software had zero support and we had to do 100% of everything ourselves if it needed changing - The operational cost savings alone covered a lot of the considerable bill. System/38 ran on 3 phase, required a fully air-conditioned space to operate in, had no feasible UPS capability and a simple power cut during business hours took roughly 12 hours to get 100% availability back from. And when things did break (and believe me, they did) then parts were increasingly scarce and expensive. - It was literally impossible to get any disaster recovery as no one else used that crap anymore (and a disaster did occur a while after I left when someone literally burnt the offices down, yet having pivoted to current kit meant the business was going again within a day and is still going now 25 years later) - It was quite hard to recruit staff as no one really fancied working on hardware/software that was almost as old as they were And I could go on for hours about how much better it is to write software now versus on those kinds of platforms. An awful lot of things we just take for granted now (testing frameworks, source control etc) just weren't really a thing on that vintage of platform. Things have moved forward with bloody good reason.

Except for DCS then the base product is free. Nor is it likely the case that any/every other DLC will work with DCS 3.0 without any change. The problem is that the platform is moving on and the only DLC that absolutely can't move with it is the Razbam stuff because no one who can is currently willing to change it to keep up. I used to jokingly refer to Windows 7 as Vista Service Pack 2 (I didn't think Vista was actually as bad as the press it got once 3rd party driver support actually caught up with it).

And quite a few got royally bitten on the backside for continuing to use it. The UK NHS for example (who were also paying a substantial sum for some level of continuing support after it was originally retired by Microsoft).

Century old locomotives that are basically a labour of love and rely on armies of volunteers and donations to be kept running for the most part. There is very good reason that those locomotives are now all on heritage railways or running the occasional enthusiast mainline day trip and not in day to day use. What Aapje describes is pretty much spot on. Software often can and does eventually become uneconomic to maintain and/or unsupportable through a multitude of reasons. Sometimes sustainability is throwing something in the bin after you've come with a more efficient replacement.

Fortunately those modules are in the hands of active developers who have the means and motivation to address whatever breaking change is going to come up. The Razbam ones are not.

I would suggest its eventually inevitable. They've deprecated stuff before and taken it off sale to new users (albeit after offering upgrades to actively supported modules for a fairly nominal fee) and one day even just keeping the lights on for those deprecated legacy modules won't be worth it any longer to them either.

No one said it did wear out. Maintaining backward compatibility can be a huge millstone though and its unfortunate that those modules are reaching a tipping point of not being able to be supported in their current state and can't be improved to a new standard either.

Yes, a thawing of relations and a resumption of support/development solves a lot of problems. I've not (yet) cashed in my Strike Eagle for credit in the same slim hope that eventually everyone sees sense. If that doesn't happen I accept the world needs to move on (and I do have 2 RB titles in my locker so I do have a little skin in the game if they get completely dropped). I suspect we're of a similar age as I played all those titles. A much simpler time where the number of patches a game received varied between 1 and 0 and games didn't have to ring home to licencing servers every time you started them. It felt like you owned it, not just licenced the right to use it. Its a different world now though. Ecosystem titles like DCS are only as secure as the sum of the ecosystem. Contributing developers can go bust, lose interest, fall out or just move onto other things. Even ED themselves could go bump one day and the licencing servers go with it and it goes dark for everyone. So in that sense and in this age, assuming any or all parts of it will be around forever is not a wise assumption for anyone to make.

Yes, Il2 GB has been around a while (mid 2010s I think?). Like you say though, single studio. And that is still a DX11 title and has been for as long as I remember (and it sounds like Vulkan will be the breaking change for the RB modules without maintenance). MS2020 to MS2024 then yes, a decent chunk of DLC was upgraded for free/nominal amounts but that was very much an outlier. I've had pretty much every version of MSFS between 98 and 2020 (2024 seemed too much of a bin fire on release to tempt me to get it yet) and its generally been a pretty clean break with maybe the very occasional discount if you had a given piece of DLC for the previous version.

The DLCs in question aren't for sale any longer though.

The line between 'new title' and 'new version' can be pretty blurred. The changes being banded around that are coming when we get past 2.9.x are pretty fundamental. Would it feel any better if it was DCS 2025 1.0 rather than DCS 3.0? And by the sounds of it then none of the other modules will 'just work' across that line but will likely require some level of change to keep them working. Except in the case of the Razbam modules then no one can work on them right now. So in light of that then do ED just postpone any changes that will create a breaking change in any module in its current state?

Yes they quite possibly would've all been maintained for 3.0 in the alternate reality where RB and ED never fell out. Although whether a publisher can eventually say 'enough is enough and we're sunsetting this 7-8 year old module now' I don't know. DLC being maintained in perpetuity isn't a thing on any other game i've ever played, and whilst I certainly expect a decent run out of DCS modules I don't expect them to work on the latest core DCS product forever. And longevity of modules on DCS is actually remarkably good e.g. the Harrier has been around since late 2017 and the original FF A-10 even longer. We're only at risk of losing the RB Four due to extraordinary circumstances. However the only people who can do that maintenance work to get the modules beyond 2.9.x are no longer working on DCS. So what are ED meant to do? Limit any enhancements to core DCS to those that won't break the 4 modules that DCS have no control over?

Yes there are compromises, much like there are if you choose to keep using Windows XP. And whilst Microsoft has tried to maintain backwards compatibility, there is always a point where enough is enough and they cut stuff loose. e.g. try opening a 16 bit Windows exe on Windows 11. Yeah I play AC. I have the original, I have Competizione and I also have Evo. All have their own DLC that isn't cross-title compatible even though there is significant overlap. I don't quite buy into the 'DLCs are much cheaper' slant. Some are cheaper because they're a damn sight less complex than a FF DCS module. If we're talking about DLC for other titles that is of a remotely comparable 'study level' depth then i've paid more for some MSFS titles (e.g. PMDG aircraft used to be eye-wateringly expensive but also very good) than I ever have for a DCS module (and i've bought quite a few of those at launch price). 'Long term' in MSFS is about half a dozen years. The RB Harrier in DCS has been around for 7.

Can you name me a comparable software title where DLC is bought once and then maintained in perpetuity across major releases? As others have mentioned, MSFS you basically stuck with the old version or have to start over on DLC every 4 years or so. Most racing sims i've played you end up having to start over too on DLC when a new full release comes along if you want to move up with it. And this isn't even a relationship breakdown situation like the ED v RB spat, its just what happens every few years.

The really funny thing is that if you did similar and stuck to an old build of DCS you might be able to still fly the Hawk and you quite probably will be able to still fly the Razbam modules if you forgo updates beyond 2.9.

You're right, we don't. No evidence work hadn't resumed and no evidence it did either. And given RBs previous comms style of trying to steer the narrative then I'd be astounded if they had some work ready to ship and hadn't mentioned it.

Yes it might get sold off and taken on as a going concern by someone else but it isn't an absolute given.

You'd potentially have no licence servers though if they went under. So it'd be dead to everyone in a week or whatever the maximum offline period is.

Would you be happier with a base platform that was constrained by never ever being able to cause a breaking change to a now unmaintained third party module? No one broke your car, its more the case that no one is willing to maintain the RB modules currently to keep step with the underlying DCS platform.