*** Microsoft Windows 11 Thoughts & Discussion Thread ***

It's always people who think they know better, there'll be some know-it-all dev in the team saying "but there's a better / more modern way of doing this" and then the team finds themselves spending time rewriting code that worked fine.
 
Exactly. You can't tell me that the old control panel development source code files aren't sitting on a server archive within MS that could be pulled into Visual Studio ( or whatever they use internally ), and the UI form re-done with updated looking assets. Or am I really missing something.

Just making it 'look' coherent throughout the operating system would have helped Win 11.

Everything feels like its being dumbed down.
 
The underlying code is definitely still in existence - 1-2 people like Dave's Garage have had permission to do videos on some of it - but it is using old school Win32 programming C++ but using mostly C conventions and very fragmented in approach, etc.

That said it isn't some crazy task to go through it and pull out the core functionality and build a decent control panel on a more modern platform and style - ironically the harder part is probably wrestling with the moronic interfaces and methods used in more modern platforms like UWP to reimplement.
 
Donnie Fisher said:
Exactly. You can't tell me that the old control panel development source code files aren't sitting on a server archive within MS that could be pulled into Visual Studio ( or whatever they use internally ), and the UI form re-done with updated looking assets. Or am I really missing something.

Just making it 'look' coherent throughout the operating system would have helped Win 11.

Everything feels like its being dumbed down.
Click to expand...

There speaks someone who has never tried development :)

The issue is essentially what developers would term 'spaghetti code'
When the underlying basis of your codebase is 30-odd years old, and has been added to piecemeal over the years, it becomes nothing short of a nightmare to work on.

A complete re-write was in theory a fundamentally good idea.
It's in practice where it falls down- Over those 30 years, things have been bodged on to make the fantastic new feature of the day work, using workarounds which quite often have been lost to the mists of time.

Basically in terms of the pure UI, IMHO we're stuck in a catch-22. You can either keep adding to the leaning tower (and see where that gets you with all the ongoing security issues every Patch Tuesday), or throw everything out and start again.
This inevitably looses features, and by the time the 'new' application reaches feature parity it'll have exactly the same issues all over again.....
 
leezer3 said:
The issue is essentially what developers would term 'spaghetti code'
Click to expand...
I don't disagree however the spaghetti'ness of the code base was, and still is to a certain extent, a conscious choice MS made. They could've gone down a road similar to that used by Linux where the code is more containerised, where code can be changed, updated, or even removed without a risk of the whole house of cards coming down.

Where if a new feature depends on code from 30 years ago that code can be removed when there's no more features that depend on it. MS made a choice in the early days of speed of release over maintainability of their code base, a choice they're now paying the price for IMO.

e: I will add that what I'd like to see is MS do a complete code rewrite of Windows to make it more dependency orientated, where code/programs/features can be added/removed/updated without running the risk that some obscure seemingly unrelated part of the OS depends on that code. However at this stage the Windows code base is so big it would probably take years or decades, if it's even at all possible.
 
Last edited:
leezer3 said:
There speaks someone who has never tried development :)

The issue is essentially what developers would term 'spaghetti code'
When the underlying basis of your codebase is 30-odd years old, and has been added to piecemeal over the years, it becomes nothing short of a nightmare to work on.

A complete re-write was in theory a fundamentally good idea.
It's in practice where it falls down- Over those 30 years, things have been bodged on to make the fantastic new feature of the day work, using workarounds which quite often have been lost to the mists of time.

Basically in terms of the pure UI, IMHO we're stuck in a catch-22. You can either keep adding to the leaning tower (and see where that gets you with all the ongoing security issues every Patch Tuesday), or throw everything out and start again.
This inevitably looses features, and by the time the 'new' application reaches feature parity it'll have exactly the same issues all over again.....
Click to expand...

A lot of the original APIs used have been depreciated as well and a lot of coders hate producing code using depreciated functions.

Unfortunately the replacement interfaces are often horrific - trying too hard to be fancy and clever and ending up spaghetti messes themselves, long winded for simple uses and unfriendly to read. Not to mention MS's habit of getting them partly done then abandoning them for a replacement that the suffers the same fate.
 
If you think Linux is any better in terms of code mess or horrifically stupid decisions, it's really not :)
This is the utter bane of anyone trying to provide backwards compatability, and the reason the tower just keeps growing.

xkcd.com

Dependency

xkcd.com xkcd.com

The only real difference is that Linux has a wider set of choices in terms of which package provides what, and because of this and the smaller userbase you just don't get the level of fuss you do with Windows.
 
ChrisD. said:
They’re more bothered about forcing AI and ads on people over user experience.
Click to expand...
I loved that recent update that decided to change the wallpaper to some random photos, with a shortcut on the desktop to find out more...
It's trivial to change back, but why do it at all? The constant unwanted app additions, or changes, or undoing settings. People are paid money to do this. :|
 
Last edited:
leezer3 said:
A complete re-write was in theory a fundamentally good idea.
Click to expand...
I think part of me would be more accepting of them replacing the old with the new and fresh if they actually completed it. We end up with UI overhauls being started, but then never finished as new UI frameworks seem to come and go every other week.

As it currently is, the Control Panel may be obsolete appearance wise, but at least it can't get any more obsolete. By contrast we have a lot of newer features in Windows that either look dated when they launch (Windows 11 Taskbar/Start Menu isn't going to age well imo), or are outdated soon after and then abandoned (e.g. the mish mash of Win32/.NET/Metro/WFP/??? flavour of the week UI design)





I think the bigger issue is that there simply isn't the drive and ambition at Microsoft anymore. In the 80's and 90's Microsoft were competing with a lot of big companies - competing against IBM's OS/2, Unix, and even internally with the different Windows Home (9x) vs Business branches (NT).

There's no focus on testing, compatibility or even efficiency:
Dave Cutler - forced the Windows NT Development team to use the version they were developing as a "daily driver". If it wasn't stable enough for them to develop on, then there wasn't a chance it was ready for production.
www.thurrott.com

Dave Cutler Talks NT, Cairo, and More (UPDATED)

There is no person more important to the history of Windows than Dave Cutler, and yet he has almost never granted interviews with the press.
www.thurrott.com www.thurrott.com

You only have to look at things like NTFS and how few revisions there were of it, yet it continues to work well even now (with multi-TB drives, SSDs and other things that weren't even dreamed of at the time. Compared to the "new" ReFS, that's had a million revisions already, half a dozen incompatibilities, and known occurrences of data loss.

Same with things like Dave Plummer's original task manager - 85Kb! It opened instantly, and ALWAYS worked. Compared to the bloaty mess of the new one that can take a couple of seconds to open, consumes a not insubstantial amount of resources, and frequently hangs.
www.theregister.com

Ex-Microsoft maverick delves into vintage Task Manager code

Plus: Have you updated to the latest version of Bloated Fetal Sacs?
www.theregister.com www.theregister.com
 
Your issue is the tower again :)

Take the example of configuring a network connection.
The basic TCP /IP networking stack will most likely primarily date back to the early NT days or before.
On top of this, things such as ipv6, newer SMB protocols and so-on have been bolted on.
The settings for each of these will most likely be stored differently, and there will be hacks to actually make different things talk to each other- Remember these new features weren't even thought of when the initial code was written.

When you come to implement the 'new, shiny' replacement UI, the logical thing to do at this point is to clear up the mess, and store the settings etc. in a single unified place.
 
leezer3 said:
On top of this, things such as ipv6, newer SMB protocols and so-on have been bolted on.
Click to expand...
Maybe they should've considered that to begin with, maybe they should not have used a monolithic approach to programming Windows so things that get bolted on at later date don't have to depend on "hacks to actually make different things talk to each other".
 
Armageus said:
Same with things like Dave Plummer's original task manager - 85Kb! It opened instantly, and ALWAYS worked. Compared to the bloaty mess of the new one that can take a couple of seconds to open, consumes a not insubstantial amount of resources, and frequently hangs.
Click to expand...

I had a go at the old school approach recently like Dave used back in the day when coding some software to interface with a DIY UPS setup I've been tinkering with - no framework, C++ but mostly C conventions, etc. my main exe is 35KB without the runtime embedded or 210KB with (/MT). These days software of similar complexity is usually like a 15MB executable :s

It is quite interesting as the Win32 API is quite ugly in places but at the same time I quite enjoy it - there is something quite satisfying about the approach I don't get using more modern iterations of Windows development.
 
throwaway4372 said:
This distro looks pretty easy to transition to, might be the way I go eventually.
zorin.com

Zorin OS - Make your computer better.

Discover the alternative to Windows and macOS designed to make your computer faster, more powerful, secure, and privacy-respecting.
zorin.com zorin.com
Click to expand...
I use it on an old Acer SFF which I have as my test bench when reviewing connected devices and so on. Works really well on an ancient machine and has features that are rather convenient which are missing on any version of Windows.
 
Microsoft quietly makes controversial Recall feature uninstallable - Neowin
After spending a few months in complete radio silence about Recall, Microsoft finally emerged with a statement that its controversial feature will make a comeback later this year, in October, to be more precise. In preparation for the release, Microsoft quietly made a big change in Windows 11 version 24H2 on Copilot+ PCs, namely, adding the ability to uninstall Recall (via Deskmodder).

In KB5041865, which was released for Copilot+ PCs earlier this week, Microsoft introduced Recall as an optional feature. You can press Win + R and type "optionalfeatures," and then find Recall in the list of available components (assuming you are on a supported device). Toggling it off should remove Recall from the operating system.

Until KB5041865, it was only possible to turn off Recall. Now, however, Microsoft appears to be ready to give users more control over this feature, which many find quite questionable privacy-wise.
Click to expand...

I guess that'll be one of the first things I do when installing Windows 11 then.
 
You must log in or register to reply here.
Back
Top Bottom