Category Archives: software development

amazon, apple, cloud computing, mobile, multimedia, software development

Amazon entices Android developers with $50 incentive

1 Comment

Amazon is offering Android developers $50 of AWS (Amazon Web Services) credit if they submit an app to the Amazon Android app store.

Although the announcement refers to apps that actually make use of AWS, this does not seem to be a pre-condition:

September 7 – November 15: Android developers who submit an app that is approved to the Amazon Appstore for Android through October 15 will receive a $50 promotional code towards the use of AWS products and services

The move ties in with reports of Amazon developing its own Android-based tablet/Kindle. Exactly what Amazon will offer is still under wraps.

Amazon is an interesting contender in the mobile wars because it has its own instant ecosystem – millions of customers who are already signed up with accounts and stored credit card details. Add in Kindle eBooks, the MP3 store, and the Amazon Instant Video Store for streaming video, and it amounts to a comprehensive content offering that approaches that of Apple.

The AWS element is also significant, and in this respect Amazon is ahead of Apple. Of course there is nothing to stop you using AWS with apps for iOS or other platforms, though there is synergy when it comes to payments.

The relationship with Google is interesting, in that Google controls Android but Amazon is not hooking into Google services or the official Android Marketplace. Amazon is showing no sign of developing its own search engine though, so Google will still get some benefit if Amazon devices are popular, provided Google remains the default for search.

microsoft, silverlight, software development, visual studio

Windows Phone 7 apps, stats and future

2 Comments

Justin Angel, a former Microsoft employee who worked on Silverlight, has posted his analysis of the 24,505 apps he found in the Windows Phone 7 marketplace, exploiting a loophole that lets you get the download links. A few highlights:

  • 97% of the apps are not obfuscated, meaning that it is trivial (with easily available tools) to decompile the source.
  • 90% are Silverlight vs 10% XNA. This is not so much an indicator of the popularity of the two frameworks, but more an indicator of how many apps are graphic-rich games rather than some other kind of utility. Of course if you are making a very simple app, Silverlight is easier than XNA, so that may be a factor too.
  • 99% are C# vs 1% Visual Basic and a smattering of F#. A fascinating stat that makes me wonder what is the future of Visual Basic.

There are more interesting stats about libraries and components used, for which I refer you to the original post.

Does it matter? Well, Windows Phone 7 has not been a big success so far, though the reasons for that are not so much the quality of the OS or the ease of developing apps, but rather its low profile at retail and the fact that most operators and manufacturers don’t really need it: Apple and Android between them pretty much have the market.

That said, there are a few reasons why Windows Phone or some evolution of it may yet be significant. Nokia is betting on it, and while Nokia is undoubtedly in difficulties, this must work in Microsoft’s favour. Further, fear uncertainty and doubt surrounding Android patent and copyright issues may persuade some industry players to give Windows Phone another look.

Perhaps more significantly, when Microsoft unveils its developer strategy at the BUILD conference next week, it is likely that the application model in Windows Phone, or some evolution of it, will integrate with what is planned for Windows 8. NVIDIA is already talking about how Windows 8 will run Windows Phone apps.

For these reasons I believe there is at least a glimmer of hope for Microsoft in the mobile world; certainly the developer story to be officially told next week will be an interesting one.

apple, delphi, software development

Hands on with Delphi XE2 for Apple iOS

23 Comments

Last week Embarcardero released RAD Studio XE2. RAD Studio is the suite of tools based on Delphi, a language – originally called Object Pascal – and visual development tool which still has a loyal following. XE2 is the most interesting new release for years, introducing a 64-bit compiler for Windows and cross-platform support for Apple’s OSX and iOS.

I have been trying the final release, paying particular attention to the iOS support, bearing in mind the importance of Apple’s mobile platform. The RAD Studio IDE only runs on Windows, so the most convenient way to target Apple’s platform is to install on a Windows virtual machine. I used a Parallels VM running Windows 7 64-bit, hosted on OS X Lion.

Setting up for iOS development with RAD Studio XE2 involves several steps. First, you have to use the new FireMonkey application framework in order to do cross-platform work. FireMonkey emerged after Embarcadero acquired the intellectual property of a company called KSDev early in 2011, along with its founder Eugene Kryukov:

KSDev’s intellectual property has been purchased by Embarcadero Technologies, the makers of Delphi and C++Builder Rapid App Development Tools. I am excited to announce that I have joined Embaracadero’s next gen frameworks team leading a very exciting project. As a result I will no longer operate the KSDev company and will not be accepting any further orders for KSDev products.

The products in question were Delphi frameworks called VGScene and DXScene, and these seem to have been melded with remarkable speed into what is now called FireMonkey. FireMonkey controls such as buttons and listboxes are all custom drawn, which is good for cross-platform consistency, but bad if you want your application to look and feel truly native. FireMonkey is not compatible with Delphi’s VCL (Visual Component Library), though the basic controls like TButton and TEdit are similar. FireMonkey applications can be either 3D, with the emphasis on Flash-like visual effects, or HD, used for more traditional user interfaces.

Support for Mac OSX is more fully integrated than for iOS. You can easily add an OSX target to a FireMonkey application, but for iOS you have to create a new application that only targets iOS. Another difference is that Embarcadero has its own Mac compiler, whereas the iOS support depends on the FreePascal open source compiler. If you are targetting OSX, you can code and debug entirely from the Delphi IDE, whereas for iOS you have to export your project and compile in Xcode.

In order to prepare for iOS development, you first need a Mac with XCode and the iOS SDK installed. Next, install RAD Studio XE2 on Windows. Then find the FireMonkey-iOS folder in the directory where RAD Studio XE2 is installed. This contains FireMonkey-iOS.dmg. Copy this to the Mac side, mount it and run the FireMonkey iOS installers to add FreePascal and the FireMonkey libraries to your XCode setup.

image

If you are also doing OSX development you will also need to install the Platform Assistant on the Mac, but for iOS this is not required.

Now you can go over to the Windows side, start a new application observing all the tasty new options, and choose a FireMonkey HD iOS application.

image

This creates a new form sized for an iPhone 4.0, though of course you can amend this. There is a tool palette which looks well-stocked with components, but note the following warning:

While you are designing your iOS application, you can only use components that are supported on iOS devices. However, the Tool Palette might contain components that are Windows-only or otherwise not supported on iOS.

That is an annoyance, and contributes to a feeling that iOS support is a little, dare I say, unfinished. Still, undaunted I built my sample app, following the path I have trodden before by creating a simple calculator.

image

You might wonder why all the buttons are green. I did, too, and played around a little trying to change it. This seems to involve creating a custom style. I started doing this, but decided it was not necessary for my simple test. It does make the point that the default appearance does not have the iOS look and feel.

There is what seems to me a small bug in the designer. If you select more than one control, the sizing tabs disappear and there is no visual evidence that the controls are selected, other than a heading in the Object Inspector that reads “n items selected.” At first I thought it was impossible to select more than one control, but this is not the case. However, there is no clipboard support in the visual designer. For example, if you want several buttons that are exactly the same, you need to add them individually, then multi-select and set the properties as needed.

While developing an iOS app, you can test it by running it on Windows within the IDE. When it is ready to test on iOS, you need to export the project. To do this, you need a command-line tool called dpr2xcode.exe, which is in the RAD Studio bin folder. Running this from the command-line is inconvenient, so the usual approach is to use Configure Tools from the Tools menu to add it to the IDE.

image

It is puzzling that Embarcadero has not included this by default.

Running the tool creates an xcode sub-folder in your project directory, with an .xcodeproj project file along with some default icons. I then copied the entire project folder to the Mac. It is also possible to use a shared folder accessed from both Windows and Mac, though I found this does not work if the folder is on the Windows side, so I simply copied it back and forth.

I opened the project in Xcode, and was prompted to “Modernize” it in Xcode jargon, to no ill effect. At this point I could successfully build it and run in the iPhone emulator.

Of course I wanted to test it on an actual device. I attached an iPhone 4 and did the Apple provisioning dance. After the usual messing around with certificates, it worked.

image

and here it is on the iPhone:

image

It works, and to that extent I am impressed. That said, I am disappointed with the performance. This is subjective, but I am talking about the responsiveness of the UI. There are perceptible pauses, which for such a simple app is surprising. I have created this same app numerous times using different development tools, and had expected that the Delphi version would be up there with the best, but while it is acceptable it is less responsive than some of the others.

Let me add though, a Delphi developer will find the process described above a easier than learning Objective C, and I was able to create this fully working app in an afternoon so I should not complain too much.

Maybe when Embarcadero comes up with its own iOS compiler there will be some improvement.

cloud computing, google, professional, software development

Google gets serious about App Engine, ups prices

Google App Engine will be leaving preview status and becoming more expensive in the second half of September, according to an email sent to App Engine administrators:

We are updating our policies, pricing and support model to reflect its status as a fully supported Google product … almost all applications will be billed more under the new pricing.

Along with the new prices there are improvements in the support and SLA (Service Level Agreement) for paid applications. For example, Google’s High Replication Datastore will have a new 99.95% uptime SLA (Service Level Agreement).

Premier Accounts offer companies “as many applications as they need” for $500 per month plus usage fees. Otherwise it is $9.00 per app.

Free apps are now limited to a single instance and 1GB outgoing and incoming bandwidth per day, 50,000 datastore operations, and various other restrictions. The “Instance” pricing is a new model since previously paid apps were billed on the basis of CPU time per hour. Google says in the FAQ that this change removes a barrier to scaling:

Under the current model, apps that have high latency (or in other words, apps that stay resident for long periods of time without doing anything) can’t scale because doing so is cost-prohibitive to Google. This change allows developers to run any sort of application they like but pay for all of the resources that your applications use.

Having said that, the free quota remains generous and sufficient to run a useful application without charge. Google says:

We expect the majority of current active apps will still fall under the free quotas.

Free apps are limited to a single instance, 1GB outgoing and 1GB incoming bandwidth per day, and 50,000 datastore operations, among other restrictions.

The pricing is complex, and comparing prices between cloud providers such as Amazon, Microsoft and Salesforce.com is even more complex as each one has its own way of charging. My guess is that Google will aim to be at least competitive with AWS (Amazon Web Services), while Microsoft Azure and Salesforce.com seem to be more expensive in most cases.

.net, database, microsoft, professional, software development

Full circle for Microsoft database APIs as OLEDB for SQL Server is deprecated

2 Comments

Microsoft’s Eric Nelson has posted about how the OLEDB driver for SQL Server is being deprecated and will not be supported beyond “Denali”, the forthcoming version.

OLEDB was created to be the successor to ODBC – expanding the supported data sources/models to include things other than relational databases. Notably OLEDB was tightly tied to a Windows only technology (COM) whilst ODBC was not (Although we did try and take COM cross platform via partners)

ODBC never did get replaced. What actually happened is that ODBC remained the dominant of the two technologies for many scenarios – and became increasingly used on none Windows platforms and has become the de-facto industry standard for native relational data access.

ODBC was as I recall Microsoft’s first attempt at creating a universal database API.

The death of OLEDB will be slow, according to Nelson. The OLEDB driver for Denali will be supported for seven years following Denali’s release. He also says that OLEDB itself, as opposed to the SQL Server OLEDB driver, is not necessarily being deprecated; though frankly if Microsoft ceases supporting it with its own database I cannot see much future for it.

Note that ADO.NET, which to some extent replaced OLEDB, is not being deprecated. However ADO.NET is only usable from .NET applications. When you consider that Microsoft may be to some extent tilting away from .NET and towards native code, the deprecation of OLEDB becomes even more significant.

ODBC is not particularly easy to use in its raw form. However, you can wrap ODBC with, yes, an OLEDB provider or an ADO.NET provider; or you can wrap the whole lot in an object-relational framework such as Entity Framework.

One more chapter in the long, strange and tortuous history of Microsoft’s data APIs.

microsoft, software, software development, windows, windows 7

Windows Live Messenger error message hell

Recently I tried to sign into Live Messenger on Windows 7, only to be informed of what appears to be a temporary interruption in service.

image

Show details, by the way, shows Error code: 80040154

I retried and got the same message, so I clicked the Get more information link, which took me here:

image

The help document says that the solution is to reinstall Windows Live Essentials. I confirmed this, not by reinstalling (yet), but by trying Live Messenger on another machine, where I could sign into the same account successfully.

A few observations about this:

  • The error message is incorrect, since the error is apparently on the client whereas the message states it is a temporary problem with the service.
  • Microsoft’s engineers know that the error message is incorrect, since the help document references both the message and the solution.
  • The error message has been incorrect for years, since it applies to both Windows Live Messenger 2009 and 2011.
  • The misleading error message is particularly annoying, since if the user wants to use Live Messenger urgently they may well wait as advised by the message, not realising that the problem can be fixed immediately.
  • The solution is a brute-force one that involves many other applications, including Live Mail and Live Writer (on which I am typing this post). Or is it enough just to reinstall Messenger? The message suggests not. However if you use Live Mail for all your email, you probably want to know whether the uninstall will delete all your email and contacts or not. The referenced article on uninstalling Live Essentials does not say.

How could Microsoft improve this? At risk of stating the obvious:

  • Give an accurate error message.
  • Give a solution which targets the exact problem rather than relying on an uninstall/reinstall procedure that changes many things that are working fine.
  • If this is impossible, at least advise the user in one place concerning their obvious questions, such as “what happens to my stuff”.

Update: I found the cause of this problem. A developer tool beta had overwritten my system path with its own, breaking this among other things. I do not blame any application for breaking in these circumstances. I fixed the Live Messenger problem by performing a Repair on Live Essentials – less risky than uninstall/reinstall, and in this case sufficient.

Another idea if you have this kind of problem is System Restore.

Nevertheless, the error message could do with some work!

adobe, flash, internet, mobile, silverlight, software development

Adobe says role of Flex and Flash has changed, makes play for mobile

5 Comments

Adobe’s Andrew Shorten has posted on the future of Flex, the developer-oriented tool for building applications for the Flash runtime.

This is one of the clearest statements I have seen from Adobe that recognises that the role of Flash on the web is diminishing:

There are countless examples where, in the past, Flex was (rightly) selected as the only way to deliver a great user experience. Today, many of those could be built using HTML5-related technologies and delivered via the browser, and at Adobe, we will provide tooling to help designers and developers create those experiences – Edge and Muse are two such examples.

Adobe is not giving up on Flash, of course, and states that it is still the best for certain categories of application:

We firmly believe that Flex is already the best technology for building complex, high fidelity enterprise applications such as business dashboards, line of business tools, real time trading applications and desktop replacement applications.

I would add both statements are written from the perspective of application developers. The role of Flash as a video and multimedia player is a separate issue. Flash is also important in that context. There is some overlap, in that if your application includes multimedia content then Flash is correspondingly more attractive.

As an aside, it is interesting to note that this repositioning of Flash makes it not so different from Microsoft’s Silverlight: a runtime for business applications.

Adobe is focusing on a new market for Flex in mobile. This overcomes the Apple iOS problem, since you can compile a Flex application to iOS native code. Adobe promises “additional mobile development capabilities” later this year and says:

In our next major release timeframe we expect that the need to build a fully-native application will be reserved for a small number of use cases.

I agree that cross-platform mobile development is a key area and one where there is no clear winner yet. It is a good opportunity for Adobe, though there is increasing competition from the products like Appcelerator Titanium and PhoneGap.

I also think that Embarcadero’s new RAD Studio XE2 will attract interest. This tool which will be released soon does native code compilation across Windows, Mac and Apple iOS, with Android promised, using the Delphi IDE and language.

adobe, javascript, professional, software development, Web

jQuery usage soars as Adobe Flash shows slight decline

A press release from .appendTo, a company which offers jQuery-based services and training, states that “jQuery Overtakes Flash on World’s Top Websites”. I found it a curious claim insofar as jQuery is not really an alternative to Flash, though there is some limited set of graphical effects for which I guess you could use either.

I took a look at the source data from httparchive.org – note that the data at this link changes regularly. I compared the most recent stats, from August 15 2011, to the oldest available, November 15 2010, an interval of nine months. The data is based on the most visited sites based on various lists and seems to amount to between 15,000 and 20,000 URLs.

In November 2010, jQuery was found on 39% of the sites, whereas Flash was on 49%. In August 2011, the stats show jQuery on 48% of sites with Flash on 47%, hence the press release.

Other figures that caught my eye: in web servers, Microsoft IIS has moved from 21% to 20%, apache from 51% to 49%, nginx from 11% to 13%.

Google analytics is the most commonly found script, moving from 61% to 63% of these sites. The amount of data Google receives on internet traffic is remarkable.

The real story here is the ascendancy of jQuery rather than the decline of Flash. If you want your website to work on Apple’s mobile devices as well as on desktop PCs, then Flash is not an option.

Adobe does not make money from the Flash runtime, which is free. It makes money from design tools and server-side services, among other things. Although it is good for Adobe if everyone uses its Flash client, it can still succeed in an HTML 5 world.

Flash has other roles too. Adobe AIR uses the Flash runtime on desktop PCs and some smartphones, and an iOS compiler lets you build Flash apps for Apple’s iPhone and iPad.

There is also some evidence that Adobe is tilting its efforts a little more towards HTML, with products including the preview of Edge which is a motion and interaction design tool for HTML5, CSS and JavaScript.

.net, development, professional, software development, visual studio

Continuous Integration vs Continuous Delivery vs Continuous Deployment: what is the difference?

I am reading the excellent book Continuous Delivery by Jez Humble and David Farley. But what is Continuous Delivery and how does it differ from the other “continuous” development methodologies?

It helps to understand that all these methodologies spring from the Agile software development movement, and the expression Continuous Delivery is a quote from the Agile Manifesto:

Our highest priority is to satisfy the customer through early and continuous delivery of valuable software.

Now, the starting assumption is that most software projects integrate a number of smaller projects, whether from third-parties or from team members. Since these pieces are developed to some extent independently there is a risk that changes made to one piece will require modifications to another piece; hence according to Humble and Farley:

Most software developed by large teams spends a significant proportion of its development time in an unusable state.

The business of getting all the parts to work together is called integration, and if this involves serious work you need to have an integration phase where this is the sole objective. This is a bad idea for all sorts of reasons, slowing development and preventing proper testing other than at the end of these integration phases.

The solution is called Continuous Integration (CI). You have a frequent automated build that assembles all the pieces from all the teams into a working application. If the build fails, or if automated tests run against the build fail, then this is a bug that should be fixed immediately, not later in some separate phase of development.

Tools for CI include Cruise Control, Hudson, TeamCity and others; .NET developers can also configure Visual Studio Team Foundation Server for CI.

The problem with CI alone is that the development environment is not the same as the production environment. What if the CI build works and tests pass, but once deployed the application breaks or performs badly? Perhaps the development environment runs a multi-tier application with all the tiers on a single box, but when deployed onto actual multiple machines or VMs, something goes wrong. Permission problems are another common source of errors.

Continuous Delivery means that you not only build the software, but also deploy it frequently. This usually means provisioning servers, which you can automate using a tool like Puppet for Unix-like servers, or with Virtual Lab Management in a Visual Studio environment. Automation is pretty much essential for this to work. The more closely the test environment matches the production environment, the better.

Generally though, Continuous Delivery means deployment to a test environment. What about taking the next step, and deploying continuously to production? That is the methodology called Continuous Deployment. It sounds risky; but if you have a very extensive and thorough set of automated tests, then the risks are mitigated, especially as the extent of the changes in any one deployment is reduced.

Other suggestions for reducing risk include deploying to a small subset of users first, called “canary testing”; and making rollback easy.

That said, to judge by the Humble/Farley book the distinction between Continuous Delivery and Continuous Deployment is just a little blurred. The authors acknowledge that continuous deployment into production is not always a good idea. They also imply that Continuous Deployment might mean only that your application is always ready and easy to deploy into production, not that you necessarily deploy it constantly:

Your implementation should make it possible to deploy any version of your application that has made it past the automated tests into any of your environments at the push of a button, given the correct credentials.

Compliance and security are also factors that may rightly make it impossible to automate deployment to production completely.

Client-installed applications present some special difficulties which Humble and Farley discuss.

In summary then:

Continuous integration: your application always builds and passes its tests, including all the pieces from different sub-teams.

Continuous delivery: your application always builds and deploys to a test environment and passes its tests.

Continuous deployment: your application is always ready to deploy to production through a largely automated process.

Update: I received an email from Martin Fowler about this post. He refers to Jez Humble’s post on Continuous Delivery vs Continous Deployment and adds:

– I would use your definition of Continuous Deployment for Continuous Delivery

– I would change the definition of Continuous Deployment to say something like "every good build is released to production"

However, I clarified with him that if you building for a test environment but are confident that any build that passes would be OK to deploy to production, then you are still doing Continuous Delivery. In the end, while I am sure you should use Fowler and Humble’s definitions rather than mine, it seems to me a fine distinction and that if you are doing Continuous Delivery properly then the transition to Continuous Deployment is largely a matter of policy.

adobe, internet, professional, software development, web authoring

Adobe Muse: so what is wrong with Dreamweaver?

Adobe has released a preview of Muse, a new web site design tool.

My first reaction was one of be-musement. What is wrong with Dreamweaver, the excellent web design tool included in Creative Suite? Bearing in mind that there is also a simplified Dreamweaver aimed at less technical business users, called Contribute.

Here are some distinctive features of Muse:

1. It is aimed at non-coders. The catch phrase is “Design and publish HTML websites without writing code”. Muse actually hides the code. I installed Muse on a Mac, and one of the first things I looked for was View Source. I cannot find any such feature. You have to preview the page in the browser, and view the source there. That is in contrast to Dreamweaver, where the split view shows you simultaneous HTML and visual designers, and you can edit freely in either.

2. It is an Adobe AIR application. I discovered this in a bad way. It would not install for me on Windows:

image

A curious error. Luckily I am also working on a Mac right now, and there it worked fine.

image

3. It will be sold by subscription only. The FAQ answer is worth quoting in full, as it describes one of the key advantages of cloud computing:

Muse will be sold only by subscription because it will allow the Muse team to improve the product more quickly and be more responsive to your needs. Traditionally Adobe builds up a collection of new features over 12, 18 or 24 months, then makes those changes available as a major upgrade. It is anticipated that new updates of Muse will be released much more frequently, probably quarterly. New features will be made available when they’re ready, not held to be part of an annual or biannual major upgrade. This will enable us to stay on top of browser and device compatibility issues and web design trends, as well as enabling us to respond to feature requests and market changes in a much more timely fashion.

I am reminded of Project Rome, a cancelled project which was also intended to be subscription only. Rome was for desktop publishing, Muse is for web design; otherwise there are plenty of parallels.

4. Muse promotes Adobe hosting via Business Catalyst, and if you select Publish this is the sole option:

image

Of course you can also Export as HTML. Still, it looks as if Muse is intended as part of a wider initiative which will include hosting and web analytics.

5. Muse is not a Flash authoring tool. Check out the Features page. The word Flash does not appear. Nor did any hidden Flash content appear when I exported a page as HTML. My guess: there is a quiet Flash crisis at Adobe, and the company is hastening to make its tools less Flash-centric, in favour of something more cloud and HTML 5 based. I do not mean that Flash is now unimportant. It is still critical to Adobe, and after all Muse itself runs on Flash. However it is being repositioned.

A few comments. Unfortunately I’ve not yet spoken to Adobe about Muse, but the obvious question is reflected in my heading: what is wrong with Dreamweaver? To answer my own question, I can see that Dreamweaver is a demanding tool, and that Muse, while still aimed at professionals, should be easier to learn.

On the other hand, I recall many early web design tools that tried to hide the mechanics of web pages, some more successful than others, and that in the end Dreamweaver triumphed partly thanks to its easy access to the code. Some still miss HomeSite, an even more code-centric tool. What has changed now?

Needless to say, Dreamweaver is not going away, but there is clearly overlap between the two tools.

Of course non-coders do need to be involved in web site authoring, but the trend has been towards smart content management tools, such as WordPress or Drupal, which let designers and coders develop themes while making content authoring easy for contributors. Muse is taking a different line.

Watch this space though. Even on the briefest of looks, this is an impressive AIR application, and it will be interesting to see how it fits into Adobe’s evolving business strategy.

Update: Elliot Jay Stocks blogs about the code generated by Muse, which he says is poor, and his opinion that it is too much print-oriented:

warning signs are present in this public beta that suggest Muse is very much a step in the wrong direction.