December 7, 2016
Answering to pressing requests from customers and partners, RAD Studio is moving from a 6 month release cycle with one main bug-fix update to a 1-year release cycle with multiple updates including fixes and new features.
Answering to pressing requests from customers and partners, RAD Studio is moving this year from a 6 month release cycle with one main bug-fix update to a 1-year release cycle with multiple updates including fixes and new features.
As we announced in our last published roadmap, https://community.embarcadero.com/article/news/16418-product-roadmap-august-2016, we are significantly slowing down the release cycle, going back to a more or less yearly major release for the product, from the faster cycle of recent years. There are many reasons for this change, but it mostly addresses complaints from customers (and tech partners, and component vendors).
The original requirement to release more often was driven by the fast-paced change in mobile operating system, compared to the Windows world -- which is actually now moving much faster under Windows 10, but that's a separate story. This requirement still applies, but it can be fulfilled in a different way.
This change in delivery cycle and model, in fact, is tied to another change, namely the fact that update subscription is now compulsory. I know you might not see the connection, but this gives us freedom to release new features and support new versions of operating system in updates, with no negative effect to the business financials. The only caveat, of course is maintaining the largest degree of binary compatibility with existing DCU files and packages. This might not be doable for a new operating system, but it is certainly doable for VCL and Windows, which is the platform the largest projects from our customers are on.
Berlin Update 1 was borderline, with some new features like native iOS grid added to the product, but most of the focus on fixing bugs. Berlin Update 2 has been the first this release in this new direction, with new VCL controls, new IDE designers, support for Desktop Bridge, and more.
It is true that delivering the same amount of features in a non-breaking update will require us some extra work, and in some cases (like Delphi language changes) it won't even be doable. So we might have to delay some features, because of the technical limitations due of non-breaking updates. But we feel the benefit of a slower release cycle to the stability of RAD Studio and of our customer projects, and hope this will allow more customers to stay and migrate on the latest version sooner -- with a good benefit in terms of their experience.
Needless to say your feedback is critical -- and even more because this was mostly driven by customer feedback. Do you still feel the product can move in the right direction with this model? Do you feel your update subscription remains relevant? Will you be able to safe time and money while keeping up to date with RAD Studio? Or will you upgrade your projects every 2 or 3 years no matter what? Let us know.
posted by
marcocantu @ 2:25PM | 24 Comments
[0 Pending]
24 Comments
Good
A slower release cycle with non-breaking updates will
be a reason to start a subscription again, so good
decision.
Comment by B Jansen
[http://cnoc.nl]
on December 7, 14:42
Less is More - Why Less Major Releases with Non-Breaking Updates Adding Features is a Good Thing
I agree with a slower release cycle and DCU
compatibility. I would like to have seen some of the
major bug fixes such as the ability to step into a
method on an interface in Win64 builds back ported to
Seattle. I would also like to see patch updates for
smaller issues in a more timely fashion. I have built
an AutoUpdate system if you need such a mechanism.
The ability for developers to install and test
versions with minimal time investment is very
important. I would like to try mobile development
with Delphi, but at the price point I will end up
using Oxygene, or C# with Xamarin. The market
landscape has changed....
Comment by Larry Hengen
[http://www.tpersistent.com]
on December 7, 17:21
Less is More - Why Less Major Releases with Non-Breaking Updates Adding Features is a Good Thing
I agree with the "Less Major Releases", but based on the same
roadmap you mentioned, we were still expecting to see Godzilla
before 2017. We were also told that we should see the Linux
compiler Preview before the Godzilla release, but I'm afraid we will
not see anything this year, or I'm missing something?
My subscription is about to expire and will all the changes we saw of
key developers leaving Embarcadero makes me wonder how well is
Delphi progressing.
Comment by Andre on December 7, 19:13
Less is More - Why Less Major Releases with Non-Breaking Updates Adding Features is a Good Thing
- Do you still feel the product can move in the right direction with
this model? Do you feel your update subscription remains relevant?
Yes this is still relevant but IMHO bug fixes are coming too late. A 6
month cycle to get an update is way too much.
- Will you be able to save time and money while keeping up to date
with RAD Studio? Or will you upgrade your projects every 2 or 3
years no matter what? Let us know.
I try to keep my project updated with the latest RAD Studio version
(and even when they are breaking changes)
Comment by on December 7, 19:41
Less is More - Why Less Major Releases with Non-Breaking Updates Adding Features is a Good Thing
Hi Marco, I had a big winge about the 6 monthly
installs when you were in Sydney.
Now we move up to the latest Update with a minimum of
effort. Quick and easy.
If the third party component isn't ready yet, I just
rebuild it. No need to wait.
The new way is working Marco, and I love it.
Not related, but the quality of the releases is good
as well.
Your team has clearly taken a few hits, but it hasn't
hurt you.
Made you stronger ??
Al.
Comment by Alan taylor
[http://www.altor.com.au]
on December 7, 19:58
Less is More - Why Less Major Releases with Non-Breaking Updates Adding Features is a Good Thing
I have a car. I paid for the car and I find necessary to pay the "auto
insurance" every year. I never had to use my "auto insurance" and
with the money paid I might have bought another car, but I have the
tranquility of going out on the street without fear and safe.
I think with Delphi the situation is similar and it is good and
profitable to pay the annuity to be safe and up to dated.
Comment by João Bemgiz Cordeiro
[http://www.onosistemas.com.br]
on December 7, 20:04
Less is More - Why Less Major Releases with Non-Breaking Updates Adding Features is a Good Thing
I just think that the purchase value of Delphi should be lower,
especially for small companies.
Comment by João Bemgiz Cordeiro
[http://www.onosistemas.com.br]
on December 7, 20:10
Less is More - Why Less Major Releases with Non-Breaking Updates Adding Features is a Good Thing
Look at notepad++,.
sure it's smaller, but updates are rolled out when
they are important enough.
The application launches a helper, which updates the
main .exe and any .dlls
Quick, painless, and it relaunches itself when done.
The fact that Rad studio requires the user to do a
full uninstall, and reinstall,. possibly breaking many
plugins and other dependencies - in this day and age
is the main reason for the ton of complaints about the
update process.
I don't mind updating every month - and it would make
the subscription a lot more relevant than every 6-12
months (piracy yohoho),. as long as it's minimal
effort rather than a forced full reinstall.
Comment by Thomas on December 8, 07:54
Less is More - Why Less Major Releases with Non-Breaking Updates Adding Features is a Good Thing
Do you feel your update subscription remains relevant? - Yes
Will you be able to safe time and money while keeping up to date
with RAD Studio? - Yes
1-year release cycle with multiple updates including fixes and new
features - Yep
Comment by Dobrin Petkov on December 8, 08:40
Less is More - Why Less Major Releases with Non-Breaking Updates Adding Features is a Good Thing
Was recently looking for Delphi work, got 5 interviews
(yeah!) Only one company was subscribing. The others
had decided to stay on their current versions.
Comment by Gary Sanderson on December 8, 11:57
Less is More - Why Less Major Releases with Non-Breaking Updates Adding Features is a Good Thing
This is a tough issue to address. The vast majority of work I see
available is maintaining 10-15 year-old Delphi code that's nowhere
near using the latest releases of Delphi. Currently I'm working on
something that was ported from D7 to D2007 just last year (2015).
There are extremely few jobs for NEW Delphi projects I've seen,
and not a single one that advertises the use of Delphi for mobile
platforms. For mobile, vendors want you to have experience with
the native iOS and Android dev platforms, not Delphi. So there's no
ROI value to me in keeping current since nobody looking to hire
me cares about the latest releases.
It would be nice to see a commitment to adding new language
features, but there's stuff pending now for nearly 15 years that
would be nice to see, and not much more motivation to do that
today than there ever was. Places I've worked that pull forward
legacy apps onto newer platforms are loathe to refactor code at all,
let alone to use newer features of ANY kind. So I guess there's little
value to anybody in this respect.
As an aside, I spent some time learning the latest stuff in Java 8 last
year, and made some effort to find work with local companies
hoping to leverage that. To be sure, companies are hiring scads of
Java programmers, but nearly all of them are still on Java 6. One
major SaaS vendor's VP of Software admitted they have a small
team looking into moving to Java 7, but that it would be quite a
while before they venture out into the brave new world offered by
Java 8.
This tells me it isn't a Delphi thing, but just how hesitant big
software houses are to adopt new tech. Look at what Microsoft had
to do to kill off Windows XP! Even now, everywhere I turn all I see
is Windows 7 being used, with IT grumbling about being forced to
upgrade to Windows 10 far earlier than they want. Clearly, with so
few potential users even running Windows 10, let alone the newest
version of it, what's the point of keeping on the bleeding edge of a
dev platform tracking the latest Windows features when hardly
anybody can use them?
It's a really nasty game ... vendors like Embt need to go faster and
faster to keep up with platform changes that are also happening
faster and faster. But the MARKET that's USING and PAYING
FOR this technology is quite happy to remain YEARS BEHIND the
front of that nonstop wave of change.
Comment by David Schwartz on December 8, 15:36
Less is More - Why Less Major Releases with Non-Breaking Updates Adding Features is a Good Thing
About time. I have skipped versions because I didn't
have time to do the upgrade, and then waited for
components before acting. I'm using Berlin for new
work, about to install update 2, and still slowly
upgrading remaining projects from D2007 (major
refactoring) and D6 (rewrite).
Great move.
Comment by Sue King
[http://www.torrensbs.com.au]
on December 8, 21:18
Less is Better
It's good to see them going back to what I think mostly worked a few
years ago.
Now if only they would make the pricing a little less stratospheric, I
might be able to start buying upgrades again.
Comment by Rob McDonell
[http://arkangles.com]
on December 9, 08:45
Less is More - Why Less Major Releases with Non-Breaking Updates Adding Features is a Good Thing
Personally i would rather an 1.0 update every 2-3 years and regular
patches. For me, migration to a new version of Delphi represents
months of 3rd-party updates hunting, testing and re-installation,
not to mention the work to get my many projects to run and to test
them. I cannot afford to do this every year, and this cost in my time
is immeasurably higher than the Delphi maintenance fees. I would
willingly pay 3-4 times more in annual fees to get a more stable
development tool, one whose updates were not such a time sink.
Comment by Olivier Beltrami
[www.qppstudio.net]
on December 15, 13:56
Less is More - Why Less Major Releases with Non-Breaking Updates Adding Features is a Good Thing
Stratospheric indeed. Delphi Professional is two
months rent. I'd love a version with fully featured IDE,
32bit only and some way of learning FireMonkey but
without deployment so I can do my hobby programs
and also prepare for a job.
Comment by Gary Sanderson on December 16, 07:03
Less is More - Why Less Major Releases with Non-Breaking Updates Adding Features is a Good Thing
From the pricing and positioning, I suppose that nowadays Delphi
is only about sales in Enterprise, so that limits you to stories about
grand releases that your sales people you can pitch at CIOs.
If you would like to recapture the self-serve market as to greatly
increase the user base to help raise product quality, I would suggest
to go full subscription. Basically you have to organize it in a way
that the IDE works only if you have an active suscription and the
compiler works always, even if you never had a subscription. You
just have to beat everyone else at the IDE who tries to build an
alternative tooling based on your free compilers.
If you price that subscription at about $299/year, you will be able
to focus on improving the product at the pace that your team
allows, With a subscription product, most of your users will be at
the lalest build of your latest release, which greatly simplifies
maintenance. In fact, it is what happes with free tools where most
users are at the last version just because there is no friction to go
there.
This is our own model (we evolved from a freeware tool to a
subscription-only tool) and we could not be happier. Revenues
grow without active sales investment because the entry barrier for
users is a lot lower so there is not much need for sales effort, if at
all. On the downside, you would have to fire a lot of marketing and
sales people and resellers, because with this model and price point,
it is the product that must sell itself.
Unfortunately you cannot sell permanent licenses and full
subscription at the same time. It does not work. Users do not know
which license to choose and then they choose none. You want to
give as few options in licensing as possible to users so that you can
focus on making the product great again by not having the pressure
to figure out what to say for the next big release circus.
Comment by Pep
[]
on December 16, 09:38
Less is More - Why Less Major Releases with Non-Breaking Updates Adding Features is a Good Thing
All good.
But you HAVE to sort out the forced uninstall when updating to a
new version. None of the other tools I use require 2-4 hours effort
of uninstalling, then reinstalling and sorting out the PATH etc, then
re-installing all the 3rd party components. My company charges
me out at £100/hour so we incur a real cost to having to do this.
I appreciate it may be a hard nut to crack, but the effort you have to
expend in doing so fades into insignificance compared to the
combined effort of all of us, your customers.
Comment by Niels Thomsen
[]
on December 16, 09:38
Less is More - Why Less Major Releases with Non-Breaking Updates Adding Features is a Good Thing
Agree with @Pep [] on December 16, 09:38:
Your pricing policy must change if you want to reach
more costumers.
Look what Epic does with Unreal Engine 4? Use it free
and as soon as you start making money, pay as
something back. It makes sense. This way lots of
people who normally would never look in your direction
can be drawn to your camp, start using your product
and the wheel starts rolling.
At this momment it is simply really poor strategy to
give only starter for free which is so limited.
Comment by Artur Czajkowski on December 16, 20:12
Less is More - Why Less Major Releases with Non-Breaking Updates Adding Features is a Good Thing
> Your pricing policy must change if you want to reach more
costumers.
I fully agree. In fact, our policy already changed making both
Academic licenses and Start licences free.
We plan expanding the Starter edition capabilities, to make more
features available for free to non-professional developers.
Comment by Marco Cantù
[http://twitter.com/marcocantu]
on December 18, 09:45
Less is More - Why Less Major Releases with Non-Breaking Updates Adding Features is a Good Thing
I am using Rad Studio at 2d-Datarecording ( Delphi and
c++ builder) and do own a license/subscription for
Delphi for personal use at home. Subscription was good
when a lot was added to the product (mobile support,
Firedac, added language features, parallel library). I
hope that a lot more will be added to the product in
future. That is the best way for Idera/embarcadero to
keep customers on the subscription plan. Linux server
support will be added soon. It is something I want and
keep me on subscription up to now.Still, It would have
been nice to have the possibility to compile for 32
bit Linux to develop IoT solution with Delphi/c++
builder. Nevertheless if Idera/Embacadero want to keep
me on subscription for the next 2-3 years, they have
to add the possibility to compile with c++ builder
open source project such as OpenCV or VTK. It is
first priority for me.
Comment by Alexandre Jacquot
[http://www.2d-datarecording.com]
on December 18, 11:41
Less is More - Why Less Major Releases with Non-Breaking Updates Adding Features is a Good Thing
Have read Academic edition (what does it include
btw? is it same as Architect one?) is now free. If that is
true it is not communicated from Embarcadero's
website. Only reference to Academia I found is at
https://www.embarcadero.com/how-to-buy and just
has a contact form link saying that there are special
offers for academia
Comment by George Birbilis
[http://zoomicon.com]
on December 19, 14:41
Less is More - Why Less Major Releases with Non-Breaking Updates Adding Features is a Good Thing
Sorry Marco, but I think you're overlooking the
downsides of an annual release cycle for marketing.
It's vital these days to continuously generate news.
It's the only way to maintain and expand the
developer community. And every new release is
newsworthy. Once a year is simply not enough.
I think Delphi has a lot of valuable functionality
already. And since subscription is already
compulsory, I think there's little pressure to
introduce lots of big functionality in one single
annual release. Marketing thrives on frequent small
enhancements and you can not seriously call in-
between-releases "updates" because this sounds
like "bugfixes only". I think this is a major mistake.
From a developers point of view I understand that an
annual release-cycle sounds like "stability", but this
is not true. Only by strict testing-procedures you gain
stable releases. An annual release-cycle might even
lead to less stable releases as it becomes tempting to
postpone testing.
I would suggest to keep the frequent release-cycles.
And if this leads to too much pressure on people,
then just 1) offer less new functionality, 2) test more
and 3) enhance documentation/tutorials.
One thing that should be improved is
to "automatically" install components and tools in
every new release based on the previous installation.
(Most times one would want the same
components/tools anyway. Notice that Apple is
doing this with iOS also, so GetIt should do the
same) I think automatic installation of
components/tools will probably accommodate to
most complaints on the current release strategy.
Comment by Peter
[]
on December 20, 13:35
Less is More - Why Less Major Releases with Non-Breaking Updates Adding Features is a Good Thing
""I fully agree. In fact, our policy already changed
making both Academic licenses and Start licences free.""
Wasn't that a temporary thing?
Hm,. no seems not,.
https://www.embarcadero.com/products/delphi/starter/promotional-download
It is somewhat hidden though,. and the free trial is
on a separate page.
Marketing wise I think this setup could be improved a bit.
Like, how about including the starter option in the
"buy now" page?
->
https://www.embarcadero.com/app-development-tools-store/delphi
With a link to the more info as well, so people can do
a comparison on a single page?
Comment by Thomas on December 22, 09:09
Less is More - Why Less Major Releases with Non-Breaking Updates Adding Features is a Good Thing
Although we maintain our subscription to keep our options open, we
are unlikely to update in less than five years - it's just too painful with
all the VCL updates required.
Comment by Steven Bliss
[http://www.LauckeMills.com.au]
on January 24, 01:20
Post Your Comment
Click
here for posting
your feedback to this blog.
There are currently 0 pending (unapproved) messages.