Why Windows 10 is bad

[Xen]

Robert Heller schreef op 03-12-2017 20:16:

> The problem with the "Factory Model" of "selling" software as a commody 
> is
> that the actual "production costs" of the software <...> is close to 
> zero
> and the R&D costs are huge, but are all in the past.
> Also software does not get "used up" <...> or "wear out"
> <...>, so there is no reason to "replace" the software with
> a new "copy" <...>.  So how do you pay your R&D staff <...>,
> once you ship the software <...>?

This not only applies to companies, also to parents.

Try to explain to people who cannot see, understand or value your work,
what you are doing with all your time ;-).

Software is an investment. Writing a book is also.

You can call it a problem, but that is the nature of things.

Right.

> Why you have them develop the next version, which MUST have features to 
> give the
> end-users a reason to buy the software *again*.  Even if the features 
> are
> senseless "eye candy" or other cruft thought up by the Marketing Dept.

Or come up with a subscription model for updates.
The idea that a customer should be entitled to all updates is based on 
the presence
of release-time bugs.

The better the software, the less this entitlement.

Where do you place the burden, your development team, or your support 
department?

Then, if bugfixes aren't required (as much) you introduce a subscription 
model for feature improvements.

This does not entitle people to a new version, but you coalesce the 
feature improvements into a new version _for new customers_.

For existing customers, it basically ends when the new version is 
released.

They have received all of the new features and you cancel their 
subscriptions.

So the subscriptions pay for the development of the new version.

After a period of time, you upgrade their installations to the new 
version (in fact).

As a reward for their loyalty. So they have purchased through their 
subscriptions, the new version.

You ensure that this sum is in fact equivalent to the new version minus 
a bit.

You give them a headstart on the new features for the next version if 
they subscribe again
which gives them say 6 months of feature updates for free.

So you get a rolling upgrade cycle for existing customers.

Etc. etc.

> *Except* that part of that is "release the source too".

For some reason it trickles into binary releases as well.

> Which means the software is peer reviewed. And is also subject to
> "many eyes" looking at the source: "With enough eyeballs, all bugs
> are shallow." And there generally isn't a dumbass "Marketing Dept."
> thinking up silly features to add in. And more attention is spent
> on actually fixing bugs and not on adding features solely for the
> sake of adding features.

Yet if you look at it from a meta view, it has the same dynamics as the 
current Microsoft thing.

The promise is that the bugs will be fixed later on, but in reality once 
you have released something, you are less willing to still work on it.

After release, attention shifts to customer relations, this is natural. 
So now you have split your attention (in open source this would be user 
support).

And because there are now many bugs, or the documentation is not so 
great, suddenly you have to spend a large part of your time on managing 
these issues.

The result is that your resources are drained, and that you don't get to 
the actual work you thought you were going to do.

The time you could have spent on improving usability is now spent on 
dealing with users who can't use it.

But this time is much less cost-effective (or efficient).

Because the nature of software is that it multiplies, but the nature of 
customer support is that it doesn't multiply so very well.

Anyway, this is my opinion I guess.