[Estimated Reading Time: < 1 minute]
Somebody appears to have got their hands on a draft version of a “What’s New” document for Delphi and C++ Builder XE3. Get it while it’s hot – it’s likely to get taken down with extreme prejudice quite soon.
Disclaimers:
- I have no idea where this came from
- I have no idea how accurate or correct it is
- However accurate or correct – or not – it certainly appears incomplete at this stage
/ul>
Just how useful this might be, with official information due to be made available quite soon, is questionable, but then again imho this is the sort of information we should have had weeks or even months ago. Especially given that some of the changes appear to (potentially) impact on existing code and it would be helpful to know what you could/should perhaps be doing to prepare for it.
But if you just can’t wait, it’s there (as of time of writing) if you are interested.
I can see no new features in Delphi itself. Bad, bad, bad 🙁
@Peter – ‘Bad, bad, bad’
I really don’t understand this attitude. XE2 was packed full of half-finished things (FMX, LiveBindings, VCL Styles), and you want yet more stuff? A conservative, solid release would do mightily IMO.
I do not use these features (FMX, LiveBindings, VCL Styles). I would expect BCC64 or a fixed floating point handling (look at the huge amount of open FP QCs) or a fixed Delphi Help or the Delphi Language in the front end or a expanded VCL Graphics support…
More is simply not possible at the moment, obviously. Fixed is better than anything else. This is required. Not more not less. XE2 but fixed is enough, of course improved from a usabilities perspective. If someone does disagree from the perspective of the concept behind FMX, then nothing will help anyway. If someone did attend one of the last webinars, I missed it but followed the discussion afterwards – of course what has been announced will come.
The perspective – get a box in fall and be happy – this will not work for the new things. It should do. Of course the Zeitgeist changed … what has been shipped has to work otherwise it’s useless. EMB is aware of many things a lot more than we think they would be.
Windows 8 (ex Metro)/ WinRT (although only on Intel) support would be a new feature anyway, especially if it could target Surface tablets on Intel processors and would make a valid release alone.
What new features would you need? Windows ARM? Android? Linux? They are much more difficult to implement, especially since they need new compilers and frameworks. And if they had switched to LLVM it’s clear their compiler dept. has no resources to work on many compilers at the same time. And the frameworks dept. is now so Firemonkey focused that they have still to figure out how to make it work on too many different paltform and devices.
Zero Delphi compiler or debugger changes? That’s a bit disappointing, if it’s true…
Well, there probably would be.
No, when toor is coming, they started picking my “compiler internal error” QCs and fixing or asking for more details.
Not few months ago opf course, when fixes could be made into XE2
no -> Now
It looks pretty clearly from the docwiki, i.e. the online documentation that hasn’t been made publicly available yet.
‘we should have had weeks or even months ago’
I wouldn’t be surprised if ‘months ago’ even the decision makers at EMBT did not know what exactly would be ‘new to XE3’, and I don’t say that as a criticism. Surely you’re not suggesting they should put out that level of detail months in advance? If features then slip, there would be blog posts written saying how ridiculous EMBT are for not being able to stick to their ‘promises’. In the the D2009 case it would have been WRT Delphi .NET, which was apparently only dropped late in the day; likewise the son-of-CLX thing for OS X that was apparently due for XE until quite late on.
IMO, regularly publishing detailed future plans would also aggravate the problem of EMBT apparently being wracked with fear about ‘is this it?’ responses, which has led to overly-ambitious projects crashing and burning (cross platform by 2010!), and half-finished features being stuffed in to what does get shipped. OTOH, they could just be more self-confident and ignore the ‘is this it?’ crap of course…
As I have said before, the problem is not the level of detail that Embarcadero provide in the information about their plans and progress, it is the acomplete and utter bsence of any such information that is the problem.
Certainly “months” ago this level of detail would not have been appropriate, but you cannot seriously suggest that it was only in the last 2-3 weeks that Embarcadero had any clue as to what they were realistically going to be able to include in XE3. As far as I know we didn’t even get the updated roadmap that was promised back in September last year.
Heck, they could even have underplayed it, so that we eventually we get more than they had led us to expect. But in the absence of any form of meaningful communication, customers are left to ferment and foster their own expectations based on rumour and speculation. Which is what leads to the “is this it ?”.
imho. ymmv.
“you cannot seriously suggest that it was only in the last 2-3 weeks”
Conversely, I don’t see what a fortnight’s advance notice will give you. For myself, the proof will be very much in the pudding – if the XE3 trial indicates FMX to be at the level I had expected it would be 12 months ago, I’ll buy, otherwise I won’t. Put another way, the fact I now know actions (which I had expected would be in a year ago) are included does not in itself give reason to purchase. If nothing else, hours wasted debugging the terrible FMX menu code in XE2 have taught me that!
I agree, two weeks advance notice is pitifully and meaninglessly short. My point – which I am sure you understood but for some reason are choosing to ignore or misinterpret – is that this level of information absolutely could have been made available weeks ago, giving us not 2 week advance notice but perhaps as much as 2, 3 or even 4 months notice. Even before that, if it was becoming clear that there would not be any Android support, preview or otherwise, in any initial XE3 release then this could also have been communicated, so that anyone who was holding out for that would know what to expect, come to terms with it, and get over their disappointment by the time XE3 was actually released.
Sure, there might be negative comment arising from such news, but that comment will be in the past by the release date so the release itself does not become blighted and over-shadowed by that negativity, and people might even come to see other things to appreciate in the release to assuage that disappointment.
As I pointed out in my other recent post, I knew back in April what Microsoft were planning for the Visual Studio release due next week. That’s 4 months “advance notice”, and I’m not even a Visual Studio user or customer !! Furthermore, the reaction and response to what Microsoft were working on gave them an opportunity to change direction, because it was apparent that their users and customers at large were not at all happy about it. The result from their customers of course, even though they still perhaps aren’t getting exactly what they want – is not “is this it?” but “THANK YOU!”.
I have to wonder what Embarcadero are paying you to defend the way they treat their customers with such disdain. Should we perhaps expect to see a “Community
SycophantEvangelist” banner appearing on your site in the near future ? 😉‘My point – which I am sure you understood but for some reason are choosing to ignore or misinterpret – is that this level of information absolutely could have been made available weeks ago’
And my (main) point is that I don’t care what EMBT promise (or say they ‘hope to achieve’), only what they deliver. Didn’t the disjunction between the promise and reality of FMX in XE2 teach you anything…?
‘Should we perhaps expect to see a “Community
SycophantEvangelist” banner appearing on your site in the near future?’Well if someone wants to bung me free licences I won’t object!
The document and information under discussion is what will be delivered. The point which seemingly hasn’t managed to get through to you yet is that this level of certainty doesn’t suddenly crystalise only in the 2-3 weeks before a release. The status quo is that we don’t get ANY information.
You seem to think the only realistic choices are between “No Information” and “Unreliable Information until the last minute”. My point, and the point that others do seem able to grasp, is that other companies manage to be far more open, communicative, and dare I say responsive, during the development cycles between releases of their products, thereby proving that it can be done.
The longer the gap, the more the uncertainty of what will actually ship from EMBT’s side. You’ve agreed that EMBT releasing the information just leaked ‘2-3 weeks ago’ would have made little difference to you. What time period would have been useful then? 2-3 months? 6 months? A year? 2-3 years…? In the XE2 case, even just one year may have led them with egg on their face, given the timing of the KSDev acquisition, which was presumably determined by blind panic when in-house efforts failed to produce a shipping product for XE.
I laid it out in my reply – if you can’t be bothered reading then there is no point continuing this “conversation”.
You’re right, I cannot understand why you think roadmaps *are* the ‘delivered’ products. The things I want are the IDE, compiler etc., not promises of what they will hopefully be. Naturally, I don’t agree with your analogies either (Windows 8 has barely changed over its public beta cycle, just like Windows 7 before it), though let’s not get into that.
I would guess a key reason for our disagreement is that I don’t have an SA contract, which means I am actually in the position of being able to evaluate XE3 before buying, whereas you have already bought it. Funny how it is nonetheless me who is the ‘sycophant’ while you are the objective observer…
When did anyone reference Windows 8 as an example of Microsoft listening to it’s developer customers?
The example referenced was Visual Studio. You know, the example far more relevant for comparison with RAD Studio ?
You seem intent not on having a conversation but on delivering an apology for Embarcadero masquerading as one.
I’m happy to see they’ve added SQLite support directly to Delphi.
Ditto. Presumably as part of dbExpress.
If true, i pass, waiting for XE 4!
It’s fairly underwhelming to say the least.
What is disturbingly absent is any mention of bug fixes to Firemonkey, the IDE, or enhancements for better iOS & OS/X support, as well as support for new OSes like Android and Linux. Looks like another year has gone by without any remarkable progess forward.
I also just checked QC and one of the most annoying bugs I reported QC# 99381 is still showing as open internally as of September last year. I will be choked if it doesn’t get fixed in this release (the 3rd or 4th since I reported it).
“Looks like another year has gone by without any remarkable progess forward.”
Huh? Last year brought Win64, OS X, and (on one level) iOS targetting. How on earth will adding yet more targets help make FMX any better, when its fundamentals are what need fixing?
Is worst, XE3 will not support iOS. Confirmed. They will release it , we hope, on Q1 2013.
I have seen “FMX.Materials” and the new System.Actions units being used in some (commercial) components, so there is some truth in the document…
Is this for real? Are those the new features???
– VCL/FMX “Metro” Desktop Application (no mobile, no WinRT support)
Developers from Help+Manual already created a “Metro” VCL component long time ago (http://www.helpandmanual.com/downloads_delphi.html) and DevExpress also released already a very nice component. This will be like Delphi’s Ribbon component, not really useful for real commercial softwares and most people will prefer to use the nicer and feature richer versions from third party companies anyway. And I will not mention the version for FireMonkey, because this is also something easy to create in FireMonkey anyway.
– C++ 64-bit for Windows (Clang-based)
Just really testing to see if this will be better or not
– More features for LiveBindings/Firemonkey
Maybe this could be useful, but just some updates.
– Support for SQLite databases in dbExpress framework
There are plenty of free components to access SQLite. Note sure if someone would be using dbExpress to access SQLite.
– Some FireMonkey 3D enhancements
Maybe Embarcadero is trying to convert Delphi into a development tool for games like Corona SDK, but than they would need to drop the price a lot for this. How is the percentage of Delphi users adding any 3D feature inside their application? 0.1 % maybe. It’s nice to play with it, but nothing else besides that. Last but not least, I believe people would prefer to use something more advanced like GLSCene.
Hopefully they are also fixing the old bugs and not only from FireMonkey (if done)
Let’s wait…
What is wrong with you people! As stated in the article it seems like a draft version, and looking at it, this seems true. “Zero Delphi compiler or debugger changes”… why the hell would they have added the heading if they weren’t going to add some information.
I hate dumb people, so please uses your brains before opening you stupid mouths!!
Anyway 😀 …. If you look more closely on the interwebs you might find 2 additional documents outlining the LiveBinding changes. These documents (including screenshots) cover the LiveBindings Designer and Wizard.
Seems there is more than just “some updates” to the live bindings. Seems like they are taking the Cocoa Binding path, allowing interfaces to be coded without the need for code. Take the simple Cocoa Browser sample App as the example.
Men, I have never seen such a negative group of people. The document is a draft with many new features and enhancements but all that you guys do is complain about everything. C´mon, lets show up some support here!
No iOS support. Start demanding it!
Complaining is easy. And sometimes venting just feels good.
Legitimate constructive criticism takes more effort.
I could give someone slack on the first if they also do the second.
As far as my “complaining” is concerned, you can consider the cited examples of openness and appropriate communication from the likes of Microsoft and Apple as constructive, since they indicate what I expect at the very least from a company when I am paying for their products.
Not excuses and broken promises.
Updated roadmap : Never arrived.
Anyone can ask for advance information : But nobody can know that you know
If you wanted to know above future releases you only had to say : Signing up for SA is a clear indication of interest in future releases (imho).
Why pick on us? Why not pick on Microsoft and Apple too? : Because (on this at least) they are getting it right and you are getting it wrong. And here is how…
Cue the excuse makers jumping on the irrelevant issues of price and citing references to not just different products but wholly and entirely different types of products.
Making excuses is also much easier than doing anything substantive to address shortcomings and concerns, or even simply admitting that yes, customers are being let down. The first step toward solving a problem is acknowledging that you have a problem, and this is the step that Embarcadero have yet to take.
You don’t offer a drink to a friend you can see has a problem with alcohol. You tell them they have a problem.
You aren’t the only one complaining. Some of it has some truth, some of it is a little ridiculous.
I’ll leave it to someone else to judge how constructive your specific comments are.
I think the problem with not publishing the “what will be released” months before release time is the yearly release cycle. With this deadline, embt don’t know for sure what could be in the release before it is ready 100% by the time of release.
While other vendor don’t have specific release cycle, so they have plenty of time improving the product until it is ready and able to publish what will be released months before release time, because they already entering the beta test phase and no new feature will be developed or added.
I think it is better embt make it a two year cycle release and release update during the time, and of course doing a better marketing to support the development cost.
With more frequent update, then it will make the customer happy to buy delphi because no need to buy new delphi each year just to have bug fixed and no need to upgrade the application each year to have it compiled with the latest update and fixes
I m fully agree! 2yr dev cycle is ok to release main features, and you can make “preview” of feature during releasing patches!
And I m waiting for native platform-specific controls, specially on iOS. And maybe LLVM based compiler can get rid of j# things of studio!)
Keep waiting! No iOS support on next release, they hope have something on Q1 of 2013. Repeat, NO iOS Support, if you have a XE2 iOS project will don’t work on XE3.
ABSOLUTELY. The 1 year cycle is clearly too rapid a pace. I can only imagine how much preparing a new release each year takes up in terms of their resources.
Metro templates are available. Does that imply WinRT can be targeted?
That’s pretty important.
Other than that, I am ready to wait for the official info. This is, after all, a wiki addressing specifics.
You can make a FireMonkey app that looks like iOS even tho it runs on Win32… something that looks like a Windows 8 app needn’t necessarily run on Win RT. Not saying that will be the case, but it certainly could be.
AFAIK a “Metro style” app has to work on WinRT, becasue it’s not just the “tile” interface, but it’s the whole subsystem they have to run it to be real “Metro style” apps.
That said WinRT is just an API. Because it is a C++ callable API through COM-like interfaces, it should be callable by Delphi as well on an Intel processor without many compiler changes.
Hope they didn’t just mimic a “tile interface”, and added WinRT support instead. The only question is: “how broad is WinRT support, and how stable”?
Yes, that’s how I understand it.
Targeting WinRT and the Metro API, not to mention dealing with the stronger isolation rules between apps, is probably not trivial.
My worry is it may even be impossible with a native-code compiler.
If it’s possible with C++ then it should be possible with Delphi. What it may need (or at least benefit from) is language extensions and enhancements to make the language (i.e. ObjectPascal) feel at home. But with the generics and RTTI fetish that is rampant at Embarcadero I doubt we will see any such thing.
Certainly targeting WinRT “properly” wouldn’t be trivial, but it is not impossible either.
Maybe they forgot how to write compilers any more, or realised too late that they fired everyone who could ?
At first It looks a little underwhelming, but when I look at it , it addresses the major criticisms on XE2 and supports the win 8 metro interface.
I think it might not go in the history as being the biggest revolution. But stabilising stuff and getting a better usability will be welcome for the whole app development in firemonkey.
With many directions now open, I think Emb should really tell its customers where the next release is going to allow them (us) to plan and budget as needed. I understand they may not have the resources to target Win8, Android and Linux at the same time, and for this very reasons they should tell “next version will target Win8 support” or something alike. Of course they will make someone disappointed, but IMHO it’s better to explain it *before* and *openly*, than letting speculation go and fire up disappointment exactly when the new release is unveiled. Unless they think expectations and speculations are good to bring more SAs in, but it’s a dangerous approach that is gonna to backfire since they opened the pandora box of multiplatform support.
Honestly Luigi you have this information. You get it in fall. Don’t get me wrong, of course a clear announcement maybe little understated can help but will not change a lot. The customer’s ‘false’ assumption is simply that the box shipped in fall is the final result – this is the next step in the road map.
Delphi has ‘always’ been ‘2 years’ late in the very detail, remember ActiveX these days – to foundation had been laid but a convenient access to the features had been provided in the next release in many cases, simply because EMB is not the OS vendor. With D7 a synchronization point has been introduced by accident.
Cross platform will not differ a lot from this perspective as long as the underlying operating systems evolution does continue, maybe because of other reasons for sure. In practice EMB cannot escape providing ongoing updates, they don’t know what the next version of an OS will bring. They can only try to minimize the dependencies. This is what they try to achieve. What they have to learn is to stay current with the target OS releases. The argument I have read somewhere Andorid 3.2 would be standard – this is irrelevant – the one customer who is on V4.x is the measure.
The native/native option does exist anyway provided by the device or OS vendor in a maybe more or maybe less convenient fashion.
When we think back, what Microsoft announced one year ago and what will become available, day and night – more than disappointing. Many people’s simplified view live with, what VS does provide would be the standard – the assumption behind is that MS wanted it to be this way, because of a certain higher goal or wisdom… Jolyon – VS is not the point concerning MS – it’s what they promised for Win8 on the platform level.
Let me say EMB should not ‘boast’ – the opinion of a product manager is worth nothing in general, it’s just an opinion and by a product manager.
Delphi had to be late in the past because MS didn’t release early preview “releases” and documentation as they are doing now. Sure, when you get the preview release MS is already far into the development, and VS will be as well. But Delphi can be far less late today and with better support than it had to be fifteen years ago. I would understand if I get full “Metro” support a little later than VS, but I really need to know what direction Delphi will take as soon as possibile because I have my own applications to plan and release. Waiting for the final product to be unveiled just require our to delay those decisions – and it could be a risk – with no gain for Embarcadero – if Delphi suits our needs we buy it, if he doesn’t we don’t buy it, and nothing changes if we decided it six months before or not. Just if we can’t plan in time it can impact our business as well, and in turn Emb too – because if we lose customers and revenues we have to shrink our expenses as well – and usually it starts at the most expensive end…
Delphi is 95% VCL and 5% FMX. Nothing happened for the 95% VCL people in XE3.
Considering it’s not an “official document’ I’ll wait for XE3 to find out what’s new and what’s fixed. But, it sure is fun to speculate.
I think I’ll get this version. I skipped XE2 because I’m not into cross-platform development yet, but soon I will.
Also, “actions for firemonkey” – that looks like a good “fix” to me.
By the way, it looks like they are going to rename quite a lot of UI elements – “Metro” is not an official name anymore, as far as I know.
Will we get update 5 for xe2? too many bugs in compiler, linker and vcl styles
As far as I can see so far, “Win 8 app” is to “Metro” as “XE3” is to “XE2 Update 5”
That’s a good summary for XE3! 😉
Anyone noticed the linux option in Display Preferences on the docwiki site?
I might have missed it but seems to be a new option to me.
Please confirm, whether XE3 is going to support only metro like UI or real winRT with FireMonkey UI(No matter if it is not native metro controls)
What i meat by real winrt with FM is application similar to visual C++ DirectX metro apps.(ARM version when Delphi compiler is ready, no mater if it is not ready)
If they are only supporting it by UI(ordinary windows application with Metro like UI) they are trying to misguide us and it is better to stop using EMB products.
please confirm in which way the are going to supporting metro.
I have it from a good source that there is no support for WinRT in the initial release of XE3. You get only the ability to create a Metro-like interface using either VCL or FireMonkey. The EXE produced is a WIN32 executable. There is no projection of the WinRT system for Delphi. Which means that you cannot produce an approved Metro-style app which can be submitted to the app store.
However, the latest version of Prism does produce Metro-style apps so maybe that will be their reasoning for the claim of Windows 8 support!
But yes, it is a bit misleading.
That is what I feared. Oh, well..
@Chris,
What I meant by “Looks like another year has gone by without any remarkable progess forward.” was that this year there doesn’t appear to be significant advancements, as has happened in the past. Last year was an exception with the release of Firemonkey and XPlatform support, which is one of the primary reasons I upgraded.
Of course this is complete speculation based upon a document of questionable origin, and validity. I will reserve judgement until the world tour, and official documents get released, but it doesn’t look like a version that’s worth paying an upgrade for at this point.
I also agree with Joylon that EMBT could do a much better job communicating with their customers, responding to their needs and managing their expectations.
contrary, now they can leave “Metro” moniker without fearing of litigation form Microsoft in Chris Ghisler pattern.
no ARM Compiler, no Linux support ? It’s not a new version, it’s a XE2 Update…
Is a downgrade! no iOS support!
calm down everyone! enjoy summer and just take a look at XE3 when it’s ready!
Can’t you see that all these discussions stem from the very point Jolyon Smith is trying to make? The total lack of reliable informations from Embarcadero makes its customer totallu unware of what they have to expect and how they will be able to tackle upcoming platforms, even on what is still the main Delphi platform. Emb has still a mindset strongly rooted in the ’80s, and it is unable to abandon it. All other issues come from there.
I am dissapointed too.
No Facebook support directly from IDE ?!?!
That was good troll :))
Judging from the replies here I’d say that the biggest problem facing Delphi is Delphi developers.
Which appears to be the mindset of Embarcadero too, given that they treat us as a whole as ‘The Enemy’. Hiding their plans from us, only facing us on their terms (‘World Tour’).
That’s the real problem.
Right. Something is wrong about it. They do need a better view around the developer’s requires. For example, what about the support for Linux at the side server ?. We do need to be able to use Linux as server. In my opinion, it’s the most important point since a long time ago. It’s critical because usually the expensive solutions around Windows Server are the main problem to get clients for our applications. I won’t update my old Delphi . According to these news, I believe I will be still waiting for Linux support a long time. Really a very bad news