Not all softwares can be delivered as Services

Steve Ballmer of Microsoft says that Software is becoming a service. While we heard of this cry for a lot time, I am not convinced that all softwares are suitable as services.

Whether a piece of software can be delivered as a service depends on what the software do. If you are a stock broker then you must subscribe to real time stockquote services, telephony software requires services, but writing a letter on a software service does not make as much sense.

“Software as a service” is a wet dream for business. The obvious one is it turn a lot of existing customers from “one time” or “occationally” to a constance revenue stream. When you buy software as service, you need to keep up with your (monetary) subscription or risk losing the service and more importantly, the data you store on those services.

Perhaps more importantly, like Mobile Operators in Developed Countries choosing to subsidise mobile phones to lock you into their network for a year or two in the hope of selling you other related stuff,  they can attempt to “sell” other business to you, preferably those that you did not think you need. Take for example, Gmail. Lately I see I have a small blue banner advertising stuff on my Gmail screen. Google obviously hope that while I use my Gmail service, they can tempt me to click on those advertisement to generate another, probably to them the “real” business.

Finally, a lot of software are extremely expensive, and a lot of vendor finds that open source gives them a lot of pressure to reduce cost, thus, Software as a service is attractive as it allows them to deflect costs away from end users. In essence, it is another business model to explore.

There are things I will fight against delivery as service. These includes Office Stuff such as WordProcessing, Spreadsheet, presentation software etc. These things can be solitary efforts, solitary works need no services so why go services?

Design Decisions are never ending debates….

So far, all my knowledge of OpenDocumentFormat (ODF) vs OpenXML shows that I have more information on the merits of ODF and criticism of OpenXML. Hence, it is rather refreshing to see Stephen McGibbon ripping into ODF in his blog. I do hope I can find others who are willing to do the same.

With either camps, one do not need to agree with their arguements or mud slinging, if any. The great thing is that both camps expose, analyze (correctly or incorrectly, but that is not the point) and criticise decisions making on the other camp. Filtering through these blog posts are educational for anyone who wants to see how standard writings works.

So far, the best impartial comparison is the wikipedia article on the topic.

The thing is, designing a document standard or any standard for that matter, requires a lot of design decisions. The design decisions are most needed in the “grey area”, this means endless debates, even in the design committee itself. If even the design committee has to compromise, just imagine how much ammo design decisions give to detractors.

[Aside: What triggered this blog post? A sequence of events of course. I listen to RedMonk’s podcast about the OpenXML translator project, which give a possible reason (Direct mapping to existing binary document format) for OpenXML non-mixed content approach (potentially easier for machine to parse, but not easy for human to read) as discussed in this GrokLaw article. Then it is McGibbon’s rumbling about (the lack of) SpreadSheet Formula in ODF. Today, I finally read Rob Weir’s article musing on OpenXML resurrecting a dead cat known as Vector Markup Language and how this bulged the OpenXML specification by 200+ pages.]

Office 2007 ODf Open/Save is technically “Import/Export”, not “Native Support”, but may just be good enough for Massachusett

Microsoft’s Stephen McGibbon is upset (article is in Dutch, but you can babelfish “translate webpage” it ) over Groklaw’s and Rob Weir’s comments on Office 2007’s support of Open Document Format.

The only accussation that (barely) holds water is that PJ of Groklaw should had mentioned that Rob Weir is working for IBM, a supporter of ODF. I am quite suprise that PJ forgotten to mention it, given her usually high quality writing. Other than that, a cursory analysis will show that PJ wants “native support”, something that a lot of other people, Massachusetts for example, also ask for but Gibbon is simply discussing about “basic” support for ODF. Two different goals with similar, but distinct effect.

McGibbon’s webpost of how ODF support might look like in Office 2007 demonstrates that ODF is put into a category of its own and not part of the “Save As”. Is this distinction important? Yes. Crucially, if you click “Ctrl-S” or select File->Save, you cannot save as ODF. Rather, you have to explicitly tell Office 2007 to save in ODF. It is more like an “Import/Export” function. McGibbon’s arguement that ODF is treated the same as XPS (Microsoft’s rival to PDF) and PDF shows that they choose to put them in the same category when they are, at present, not. XPS, as far as I can tell, and PDF are biased extremely heavily as document output format only, unlike ODF which is a working format for Office documents.

In my view, this fails “Native Support” tests. However, it might just be good enough for Massachusett.

There is one thing I like about the screen mockout of ODF read/write function. It is easy to spot. However, I will trade this for native support anytime, anyday.

As discuss in a previous post, an educated guess suggests that it is possible to save natively in ODF. It is the question of openning up this API. In fact, I even went on to suggest that to protect Microsoft’s IP, Office can provide the document data as OpenXML and requires ODF and other data format converter to supply data s OpenXML. This will fit perfectly into “native support” request of any tender/contract and remove any accusation of Microsoft treating others as “second class citizen”. My babelfish translation is a bit off, but the end of the article seems to suggest that one can write directly from ODF to the Office Object Model, i.e., writing it directly into Office native representation of the document, suggesting my suggestion of using OpenXML as intermediary is moot. This also fits what Brian Jones said about hijacking “Save Events” as when you hijack this event to provide your own “save”, you will have to read and understand Office native representation. The point of my post is that the need to play rough and “hijack” events is a higher hoop to jump through. Hijacking “Save Events” this way is a brute-force crack and does not play nice with other components of Office. Users is likely to believe that their office 2007 is behaving strangely It is not really necessary in the light that “Save” to other inferior formats is already provided. I do presuppose that an equivalent “Open Event” is available for hijack.

One other thing, Brian Jones finally clear up a fuzzy area about ODF support in other office product in this blog entry. Microsoft, so far, only promise a OpenXML converters for older office products. Reading between the lines suggest that one will have to run the translator separately from older Office products.

A Glimpse into the world of Windows Genuine (Dis)advantage

I know Windows Genuine (Dis)advantage cause problems to legit users. Microsoft knows it. It did not come as a surprise to me that it caught up with Paul Thurrott, a strong Windows advocate. Using phrases from a respondent to UK’s Parliament All Party Internet Group’s DRM inquiry, the market will find an equilibrium somewhere between two unworkable situations, i.e., no validation of Windows (which customers wants as it eliminates one’s headache) and totally schizophrenic validation (which the laws allows).

In the middle of the two, there will be “collateral damage” to Genuine Windows users who are wrongly labelled as pirates. Since Windows is ubiquitius, even 0.0001% means a very significant portion of users affected. We all generally goes along our merry way when everything is fine but complain fervently when things do not. Hence, it is not surprise that the internet is full of these types of story. What is lack, however, are stories that resemble real life IT deployment problem.

As with everything on the net, it is simply a matter of time. Ed Foster of InfoWorld has one now. His description of his reader’s IT department troubles is slightly alarming. PCs from a major worldwide vendor affected is not immune. Being treated like a volleyball between two companies’ customers service/technical support. The very tempting offer, especially at time of crisis, justified by the phrase “damage control”, to simply pay your way out of it by buying unnecessary licenses. Not so helpful is the fact that somebody in customer services/technical support of Microsoft decided to maintain ALL computers designated to do a particular task (issuing reactivation code) at once and at the weekend, the very time that affected companies’ technical support normally use for repairing.

Most chilling of all is we arrived at understandable situation where the customers decided that they can no longer justify the cost of applying security update as it triggers their Windows Genuine (Dis)advantage. While, leaving the machines vulnerable is not as bad as water companies decided it is cheaper not to dig up and replace leaking pipes, it does make a mockery of Windows’ Security Push. While Microsoft will says that it is the customers fault for not patching, their action is a significant contributory factor.

How important is neutrality?

How important is neutrality when it comes to collaboration? A lot.

When IBM decided to form the Eclipse Foundation to take Eclipse out of its hand, I thought it is just a manouver to placid other companies (and competitors) that wants to use Eclipse. In Dana Gartner’s podcast, one of his interviewee says that this decision is important for his company’s adoption of Eclipse as its standard platform. Still, I dismiss it. Ultimately, I thought, whoever is in charge is not important, the way the software evolves and how much the core developers listen and act on feedback is.  This is partly shaped by my experience in academic world, where there is suppose to be a “free flow” of information. More on this later.

As usual, I need a wakeup call.  It came in the form of an article talking about Microsoft’s requirement for attendees to its Windows Security Conference to sign a document that allows “Microsoft to use whatever anyone says in the conference”. The important factor to me is that this is a routine document, one that a lot of other security companies routinely signed in previous conferences of the same kind. However, this time, something changed. Microsoft is now in the security market. Suddenly, signing this document become a dilemma for security companies. There is no doubt on the importance of Microsoft organized security conference for their product. Security requires Microsoft, making of the ubiquitous Operating System, cooperate with security companies to secure the operating system. However, this time round, examplified by the document, everyone is fearing that they give Microsoft an edge over them. We see a conflict of interest here. This is a very simple scenario that even I, a pighead, can understand.

That makes me reevaluate my participation in conferences. On reexamine it, I find that I do withhold information about my work plans etc for the fear that others might “steal” my idea and most importantly, get there first! I am sure others have the same thoughts as well. Nevermind that most of the projects over here requires a lot of money, meaning that things are more-or-less in the open as others can see from your grant applications, I am going to hide what I am doing, and spring it out as a surprise 2 years later when I get there first!

To be honest, I did think that the two years gap is short enough that discussion at conferences can be considered “Open” before. Right now, I have doubts. This does not seem any shorter than the speed we see technologies appearing. The only difference is that companies normally launch it as a new product, but we launch it in conferences or as journal publication.

Hmm, interesting ….

Microsoft’s new found Principles

It emerged that Microsoft letting Google into Vista, as discussed in a previous blog post, is part of a set of 12 new tenets adopted by Microsoft.Dan Farber of ZDNet managed to distill it well.

Looking at the “Windows Principles“, the name Microsoft gives to this set of tenets, it is clear that a lot of tenets are response to external pressure, particularly Antitrust, but also Linux.

The underlying question is always, does this improve Windows competativeness? Yes. Most importantly, for the first time, manufacturers are now able to substantially customize their Windows offering (Tenets 1 to 3). While it is true that nowadays you can get computers with third parties software install, but you will find that it is effectively the same Windows Stuff plus a few applications. However, the new tenets allows, in principle, to alter the “same Windows Stuff”. I think Microsoft is taking a leaf from Linux many favours (Debian, RedHat, Suse and Ubuntu). This makes Windows interesting again.

As for tenets aimed at developers, we need to see what it means in practice. Current practices, OpenXML vs ODF and Sender ID for example, do not bore well, but things can change.

Rob Weir’s experimentation of Office 2007, Open XML and OpenDocumentFormat

Rob Weir published three blog posts on Office 2007, Open XML and OpenDocumentFormat. Here is my summary of the three posts:

• Microsoft Office 2007 “Save File” dialog and its implication for OpenDocumentFormat
• OpenXML Translator (a Microsoft-supported ODF->OpenXML converter) Performance
• Potential Problems with OpenXML’s backward compatibility and “fidelity” requirment

The three documents are eye-openners. The first questions why “hardcore” ODF supporters need to capture “Save Event” in Office 2007 to hijack that event to save in ODF format as Microsoft’s Brian Jones says ODF supporters can do. Weir successfully demonstrated that all that is required, for Office 2007 to save in other arbitary formats is for Microsoft to open up the API for “Saving Document” function. At present, Microsoft only allows saving in the formats they choose. We know there are a lot of Clever Minds working on Office 2007 and we know any programmer who attends basic software architecture 101 class is likely to design “Save as Document” fucntion to select an agent from a pool of agents and delegate the actual saving to it. Hence, what Microsoft needs to do, is to permit new (third party) agents to be included into this pool. We know it before seeing Weirs’ UI tour. Weirs UI tour simply confirms it. It also confirms that in Office 2007 at least, saving to different format is permitted. [Aside, lets bid farewell to the argument that “Save As ODF” by default should not be allowed because it loses fidelity. Save as Text File by default, which is permitted, is a few order of magnitude worse] So why the need to hijack “Save Event”? May be some retooling of the “Save Document Function”, perhaps giving third parties OpenXML document to work with is necessary to protect MS Intellectual Property. However, Brian Jone’s suggestion to hijack “Save Event” suggested at least him, if not Microsoft” believe there is no IP put at risk.

Second document is an eye openner on the quality of ODF to OpenXML conversion from OpenXML Translator project that Microsoft effectively sponsored. While it is true that we lost fidelity when converting from one format to another, OpenXML Translator translation to OpenXML format did a worse job than OpenOffice.org’s translation to binary DOC format. It is really surprising that someone with access to the full specs of both ODF and OpenXML did worse that others who have to reverse engineer DOC format. While OpenOffice.org certainly have many years practice in converting to DOC format, one must consider the fact the development team behind OpenXML Translator has been working since Oct 2005 and they had been concentrating on converting only one format (ODF’s word processor format) to another (OpenXML word processor fomat) in one direction only. OpenXML Translator’s result reminds me of StarOffice’s results when it was first made opensource

Third documents throws out interesting things about OpenXML. First it note the difference between Microsoft’s claims for OpenXML and the mandate for the ECMA committee charged with turning OpenXML into a standard. Then, more importantly, it shows that there are IP hurdles for third parties wishing to maintain fidelity with Office 2007. This is a headache for the third parties and it is possible that even Microsoft cannot help them (It might not be their IP to give away).

Good work Weir!

Microsoft let Google into Vista

Its official, Microsoft will “volunteeringly” let manufacturers set the default search engine on Vista.

Setting search engines on Vista? I must declare first that I have not downloaded the Vista Beta yet, but there is two ways I can see where this decision by Microsoft has impact: Desktop search and IE7. IE7 will be no big deal. Desktop search can be a big deal as Google and other search engines can be integrated into Explorer and other places where Microsoft put its search engines on.

While semantically correct, I wonder how much this decision is driven by the “volunteer” spirit or to starve off potential antitrust complains. To me, to “volunteer” means put something into the original design, not redesigning after receiving complaints, strong complains for that matter. Anyone still remember Google’s objection to IE7?

This triggers a thought about operating systems: Should there be a mandate that it must be open. With “Open” I do not mean “Open Source” nor “Open Standard”, but to provide a set of API for everyone to plug their product in without favouring any one vendor. After all, operating systems are application enabler. Beyond basic plumbing such as internal and external communications and resource management, everything else are application in my view. Thus Explorer/command line console are applications, CD burning and Audio programs are definitely applications. As enabler, operating systems should permits swapping one out with another, should the users or manufacturers wants to do it. I am not saying that this ability to swap things out can be done easily, but just saying operating systems vendor must reduce the required work (for third parties) as much as possible.

OpenDocumentFormat to become “MS” Standard

Yes, I am playing the acronym game here. “MS” stands for Malaysian Standard, not Microsoft.

This point was not lost on Hassan when he posted that OpenDocumentFormat is on track to become a Malaysia Standard. According to his account of the SIRIM TC4 (SIRIM is the Malaysian equivalent of NIST, and TC4 is “Technical Committee 4”) meeting, one of the many meetings ODF has to go through before being acredited as MS, Microsoft’s representative tried to delay voting by latching on the fact that ODF’s has not get the final “nod” from ISO. I am sure Microsoft’s representative knows that this is a futile attempt, but he has to try. He is duty-bound to Microsoft and his professional ethnics to play this card. It is great that the committee see through this ploy. What did surprise me is the SIRIM Secretariat representative is well briefed on the progress of ODF through ISO. Unless there are evidence to the contrary, I will assume that the SIRIM representative is neutral, not pro ODF or OpenXML (Microsoft’s alternative). Thus, it is great that a comment that can have very persuasive effect on the outcome of the voting, coming from an unbiased person. By speaking, the representative also confirmed that there is no hurdle for adoption of ODF as Malaysian Standard, at least as far as SIRIM is concerned.

Bob Sutor probably blow this completely out of proportion when cheer this as another example of ODF adoption. (Apology to Mr Sutor, rather than going to his blog directly, this time it was via a third party, my thoughts was “contaminated” by the something in the third party’s article.) If ODF becomes MS, it is just like having ISO approve ODF as ISO Standard. Nice badge but nothing more. There is no big switch of government documents to the format nor is there any pressure on private individual or companies to adopt it. The best it can do is for companies/government to remain complient with “MS” should they choose to use ODF. I see approval of ODF as MS as no more than necessary bureaucratic step to realign MS to ISO Standard as ISO Standards evolves. In fact, if ISO adopts OpenXML format as one of the standard, I will be surprise and absolutely dismay if SIRIM does not follow through and adopt OpenXML format as MS.

Another salvo in Virtualization War

Wow, so many news on virtualization in such a short time. Latest salvo from Microsoft is that it is going to partner with XenSource to allow Xen-enabled Linux to run on top of Windows hypervisor that is part of Microsoft Longhorn.

To call it a salvo, implying that there is immediate effect on virtualization war, is way off the mark. As most tech journalists has spotted, the Microsoft hypervisor in question, is at least 6 months to a year away.

Let’s dig further into the deal: Xen-enabled guest operating system will be able to run under Microsoft hypervisor. Any problem with your Xen guest OS, give Microsoft a call. If it is a problem with the hypervisor, MS will fix it, if it is a problem with Xen’s guest OS, Microsoft will forward it to Xen. Hmm…, forward report to third party…. Had we heard about something similar recently? [Googling my memory] That’s it! It is the same strategy wih the Open XML Translator! Except that in the latter, MS went out of the way to say that it is  NOT supporting the project, eventhough it contribute funding and technical help.
What XenSource gain is the cooperation from MS to get its guest OS running on MS’s hypervisor. The question that needs answers is whether does this mean Microsoft is going to twick its hypervisor to read a new image, or Xen has to twick its image to read Microsoft’s hypervisor API?

Ida Fried of CNET opined that this is the result of Microsoft’s  slow realization that Open Source is not going away. Ashlee Vance of The Register believe that this is a tactic by Microsoft using XenSource to slow VMWare down until its own hypervisor is ready. I cannot help but have this mental image of XenSource swatted after Microsoft’s achieve its objective. After a cooler mind prevail, I believe a likely event is Microsoft buys XenSource sometime in the near future.