Software application development

Since attending the Parallel’s Summit 2009 back in January I’ve been trying to listen out for virtualisation news more intently than ever. Although many companies are jumping on the bandwagon, it is the processor builders at AMD and Intel that seem to have cornered the market on snappy news angles as they talk about parallelism and concurrency in the context of server consolidation.

For themselves, Parallels appears to be selling heavily on its ‘Operations Automation’ offering, which is operations support system (OSS) for service providers who need to automate the delivery of server-based apps and other resources.

But now, I’m wondering if virtualisation is getting even more sophisticated as it spreads to every available corner of the enterprise. What may have started at the server level, has been rapidly spreading towards the desktop in the form of virtualised application delivery – and now may be headed for the high-end graphics-intensive workstation.

At least that is what the latest announcements appear to suggest. Late last night I saw news of Parallels Workstation Extreme, a ‘solution’ that the company says offers “near-native performance for resource-intensive applications”.

This then, if it works, is grown up virtualisation where support for huge memory, use of multiple CPU cores and direct access to graphics cards is not a problem as users run multiple operating systems on the same physical box.

Unsurprisingly perhaps, Parallels worked with Intel, NVIDIA and HP on its new product, which CEO Serguei Beloussov describes as suitable for users of resource-intensive applications, such as those used for oil and gas exploration.

In his official press statement Beloussov detailed the specs, “Parallels Workstation Extreme offers users support for 3D professional graphics cards via Intel VT-d and the new NVIDIA SLI Multi-OS technology, delivering near-native performance. The solution also offers up to 16 CPU cores and 64GB of RAM for guest Oss.”

I contacted Jon Collins at analyst house Freeform Dynamics for his opinion on this development as just this morning I had been reading his Virtualisation – The State of Play presentation on the web.

Collins told me that, “Breaking the emulation deadlock and enabling more of a ‘pass-through’ approach is the way that virtualisation is going in general, so this news makes sense. However while high-end graphics is an important feature for virtualisation, this is not just for high-end applications. A lack of usable graphics capability imposes a constraint which can get in the way when putting together the business case – so this should be seen as much as the removal of a bottleneck, as providing a higher-order capability for specific application types.”

Where next for virtualisation then? Well, you might like to follow that argument that suggests that the virtual machine will grow in power and dominance and start to rule other computing domains. New compute structures may then emerge as the very fabric of the IT ecosystem morphs to shape itself around the virtual data centre. Will this be a bad thing? Many would say perhaps not.

With all the talk of cloud computing, mobile widgets and rich web experiences these days, I fear some of the engine room components of the total technology stack may be getting rather less attention than they deserve.

For example, a pal of mine used to head up a company that sold storage enclosures and RAID controllers. How dull can you get right? Not necessarily, storage can be a great subject – and given the size of the planet’s ever-multiplying data store, it’s more important than ever.

Similarly, content management seems to have slipped off the techie trend scope, perhaps from a news perspective only though. A search on ZDNet.co.uk will mostly show you white papers if you go trawling for this subject.

So aiming to set the record straight. Who makes the biggest headlines in Content Management Monthly (actually, I don’t think it exists) these days? Well one of the mavericks appears to be dotCMS, it is an enterprise open source CMS that is annoying the hell out of some of the bigger proprietary players.

Talking of big proprietary players, the Gartner CMS magic quadrant is made up of IBM, EMC, Open Text, Microsoft, Autonomy/Interwoven, Vignette, Day, Hyland Software, Xerox and lastly despite its unreadable press releases, Oracle.

So that’s the open source rebel alliance and the usual suspects covered, but the company that caught my eye most during a casual web search for CMS news this last weekend (yes, I know I need to get out more) was UXA Technologies Limited, who trade under the name 10CMS.

The 10CMS User eXperience Application platform is web focused offering designed to improve site performance metrics by managing image and video content for a more, wait for it, “compelling user experience”. Shell recently bought into this product so that its content editors could deploy it and (this is the good bit) build a “sophisticated homepage that drives navigation through the global site.”

I like that. Web design news arguably suffers from an overly biased focus on the front end. This is key back end stuff, for my money anyway.

10CMS has (not a term you hear all the time) multi-tenanted architecture so that it works with other platforms; and the company appears to be very focused on monetising its ventures; as such it works closely with its client’s e-commerce frameworks.

How does it work? “Editing 10CMS UXAs requires no technical skills at all, making it the ideal medium for online marketers and e-commerce merchandisers as they can update their content as often as they require without involving a third party designer or developer,” said Fergal O’Mullane, CEO at 10CMS.

Content management is, in fact, way further over on the sexy tech scale than you might have imagined. As a pure play application development consideration, its proximity to web development is extremely close. I mentioned it in passing to Mark Fraser, managing director of Green Jersey Web Design Ltd this weekend and he gave me the below quote…

“We build quite a few open source CMS web sites for SMEs. Apart from design and development, we often spend plenty of time helping them establish good editorial and publishing practices. Without these, they simply can’t exploit the power of their CMS. It’s remarkable how many organisations spend a lot of money on CMS technology but little on harnessing it. Whichever platform is used, the human qualities of intelligence, creativity and discipline are needed to feed it.”

So if you are still awake – and I do hope you are – perhaps content management systems deserve just a few more headlines now and again. At least with this one I feel I’ve done my part.

In my ongoing quest to examine web developer and web designer touch points, I have been getting my hands pleasantly dirty with both Quark’s latest iteration and the all-consuming box that is Creative Suite 4 (CS4).

Both the new products have been considerably upgraded from their previous versions and while CS4 is a really well integrated product inside the Adobe suite; Quark has also come on leaps and bounds with a new appreciation for publishing right through from the printed page to the web.

But, in some senses, what they both appear to suffer from are compatibility issues arising from these significant upgrades.

With Quark, I am constantly exporting back to keep project settings for designers that I work with who are running older versions.

With Adobe, I am facing an interminable battle with rendering issues in the open source ‘simple’ document viewer Evince. “Evince is a document viewer for multiple document formats. It currently supports pdf, postscript, djvu, tiff and dvi,” says the web site.

Will it seamlessly work with all image rendering from InDesign PDF exports? Not for me it wouldn’t. Why was I trying to read PDF’s outside of Acrobat you may ask – well, I work with a lot of developers and DBAs who like to run Ubuntu and prefer to opt for an open source reader as well. So I’m all about ‘compatibility’ everywhere.

Quark recently stated that its QuarkXPress 8product was “independently” reported as having better Flash capability than InDesign CS4. This was a claim that I put to Adobe’s Paul Burnett who is senior worldwide evangelist for creative solutions when I attended a really good CS4 ‘Masterclass’ (I’m not worthy) early this March. You can imagine his comments without me stating them.

Remember please, at its core I was approaching this topic from a developer perspective keeping in mind that web designers can now get their hands on SDKs and hosted development environments that are vastly improved compared to what was available even five years ago.

Thinking again about the web developer-designer connection; also bear in mind the that proposition Adobe is making with Flash Catalyst, which as many will know is a design tool for rapidly creating application interfaces and interactive content without coding. Who needs the developer with this in place then?

That’s an overstatement; of course it doesn’t eradicate the need for web developers. But there is an increasing amount of automation out there isn’t there? So, is “automation” always a good thing? Especially when we have the creative nature of web design to consider.

This same idea is also seen in widgets for mobile development based on web code rather than on a specialist programming language. Both professional and hobbyist developers are now able to design these add ons – but are there any dangers and will they arise in further compatibility issues I wonder?

I asked a fellow journalist for an opinion and spoke to Andrew "Spode" Miller, founder of thinkabouttech.com who told me, “I send all my invoices and documents in read-only PDF format because I know it will render the same way on every platform, thanks to the (recently) open nature of the document format. I never get any issues with other people reading documents I've exported from OpenOffice, yet strangely, I've had odd errors, especially with printing, when using PDF files exported from software such as InDesign, even when using the official Adobe Reader on Windows! Ironically, at this years Linux Expo, anyone printing their tickets using an Open Source document reader didn't get the barcode that was needed for entry.”

Seeking divine inspiration (well, corporate inspiration perhaps) on this subject, I did track down one of Adobe’s lead Acrobat guys who helped me examine the differences between what their own technology will render and the performance of open source alternatives.

The bottom line is that if you are running Acrobat 9.1.0 (and I suppose there is no reason why you wouldn’t be) then there are some easy to use conversion tools under the ‘Preflight’ option. I say easy to use because I did use them and they did work.

The open source alternatives are faster for sure. But then so is a car if you rip all the seats out and take the doors off. Trouble is, the ride is probably not as comfortable though is it?

Anyway, I shall continue to use all the products I can so that I, personally, can remain as humanly compatible with everyone I work with.

During last year’s JavaOne event I filed a short report detailing the goings on during Sun’s pre-show CommunityOne day.

Almost a year on, the company has just wrapped up another CommunityOne developer event where it showcased the Sun Open Cloud Platform, which includes a preview of its plans to offer public clouds targeted at developers, student and startups. So is this purely programmer altruism born out of a desire to “give” to the community? Or is strategic product development aligned to profit making?

Relying heavily on Sun’s own technologies (Java, MySQL, OpenSolaris and Open Storage) at its backbone, the company’s top brass has said that the architecture behind its cloud platform is designed to create a world with many clouds that are both open and interoperable from a developer perspective. Freedom of speech in the cloud is surely no bad thing; but within all this talk of interoperability, there is no mention of standards or controls. One has to sit back and wonder for a second surely?

Alongside the general preview of this initiative for developers comes the release of a core set of Open APIs published under the Creative Commons license for public review and comment. The intention being that others building public and private clouds can design them for compatibility with the Sun Cloud.

Ah I see, we need to be compatible with the Sun Cloud – is Sun trying to say that they set the benchmark and standard in this space? They’re a big company with great technology so there’s no reason why they shouldn’t aim to set the bar. But I can’t help feeling that they just held back from saying so.

Whether this a purely philanthropic move or one being taken to increase the proximity of Sun technology to the wider programming world is perhaps open to debate.

Sun says that developers will be able to deploy applications to the Sun Cloud more or less instantaneously by using pre-packaged VMIs (virtual machine images) of Sun's open source software. Via this method, the proposition is that there is no longer need to download, install and configure infrastructure software. You can read more about this here if you wish.

Available this summer (check your diary dates for JavaOne ’09 if you want to make a guess on what Jonathan Schwartz will be announcing in his keynote), the core of the Sun cloud is driven by a storage and a “compute” service that together are hoped to deliver the combined benefits of open source and cloud computing.

Nobody yet seems to have tried to write a creative headline on this topic along the lines of – Sun brightens the forecast for cloud computing. Perhaps best that I didn’t go there either isn’t it?

With the UK’s Infosecurity Europe show little more than a month away now, the security vendor community is busily polishing up its latest batch of cure-alls and wonder-tools to aid developers and security specialists in the good fight against ‘rouge’ code.

There will be talk of firewalls, new detection techniques and possibly even self-learning apps that use social networking threads to gauge threat status as viral malware starts to evidence itself across the web.

While it would be unfair to say that some of this will lean towards scaremongering, there will no doubt be some pretty creative story pitches.

A vague whisper of the kind of report that may surface next month landed (safely and without malicious intent) in my inbox earlier this week. Centred around developer ‘confusion’ over safe versus unsafe cryptographic algorithms, there is now a “manifesto” to provide programmers with an encryption check-list to ensure safer builds result at every level.

So an unsafe cryptographic algorithms guide; hogwash, hullabaloo or home-truth?

The report is available here for free download without registration. Unsurprisingly there is a security vendor behind this; in this case it is Fortify. Credit to them for at least making it free without any extra surfing on their site over and above the link I have shown here.

Before you get your self a cup of tea and get ready to phone your CTO with ground-breaking news, this is an eight page “manifesto” in 1409 words – and 318 of those are the references appendix. So it’s not exactly the security developer’s Magna Carta.

That being said, the report’s author has taken the trouble to draw an important distinction and differentiate between problems that introduce real risk to systems being developed today, as opposed to hypothetical research focused on attacks that won’t be feasible in the mainstream for years.

As always with these things, be as sceptical about the validity of the reports and tools being proffered as you about the very threats that exist on the web and inside the systems and ecosystem in which you live. That way, we all stay sharp and we all stay safe I reckon.