When you are annoyed about something, you tend to notice other related things. I have no clue how I stumbled upon this one:

http://fci.wikia.com/wiki/Campaign_for_Document_Freedom

I'm not going to debate their claims - everyone is entitled to an opinion. The one thing that got my goat, is this line:

We request people not to pollute their friends computers and the internet with unethically encoded documents.

The emphasis is mine. Like their patron saint, open source adherents are very free with the use of such terms. This is what I object to. What is unethical about any form of document encoding?

At this point, I cooled down a bit, and started examining my position. I was getting mad as hell over just one word? I looked it up on dictionary.com : http://dictionary.reference.com/browse/ethics. I request any readers that I have to please do the same - I don't think it's ethical (heh heh) to copy and paste from there.

Big whoops pour moi. Ethics, as defined by dictionary.com, is subjective. Those...people are talking about their ethics. According to their "system of moral principles" (see definitions), its unethical to encode a document using anything other than the "standard" (yes, those are quotes. Standard is subjective too) that they propose. Okay. I apologise to the good people behind that link for getting angry for no reason.

Just for the record, my "moral principals" ( see definitions) say that thrusting one's views on others using absolutist terms is wrong. So, in my subjective opinion, the entire linked article is "unethical".

I was somewhat disappointed when India voted "no" for ISO standardization of Office Open XML (hereinafter referred to as OOXML, in this post). I was even more disappointed when Microsoft, after winning the ISO vote, made a (lame, in my opinion) complaint against the Bureau of Indian Standards (hereinafter BIS);  ऐसा व्यवहार Market Leader को शोभा नही देता.

Over the past couple of days, through Guruji* and Amol**, I learned of a "minutes of meeting" document that is available on odfalliance.in, which records a meeting of the BIS on the subject of said standard. Amol has written a post on the subject, here.

When I read the document, I was at first infuriated (like Amol) and then disappointed (like Guruji). And then my now-normal sense of apathy set in. Still, after reading Amol's post, I felt the need to write some observations. Here they are. Read the document (PDF format) first.

1) The composition of the panel is very interesting. The convener, Dr. Bhattacharya, is from Sun Microsystems. Two members who have asked important, and dare one say slightly provocative questions, are from IBM and the Free Software Foundation. This is very fair - a debate over the national vote on an international standard should have all interests represented.
2) Dr. Nagarjun's position is quite millitant, as can be expected from any member of his organization. Still, this is a meeting of national importance, not a Free Software meeting. A statement like "the entire OOXML specification document should be rejected, and only the features not present in ODF should be added" should have attracted some moderation, if not disciplinary action.
3) There seems to be repeated questions about the "design goals" of OOXML. I find it somewhat disturbing that there was no direct response from the Microsoft representatives. The following document from ECMA (which owns the standard) spells it out in clear English. http://www.ecma-international.org/news/TC45_current_work/OpenXML%20White%20Paper.pdf.  This seems to have been created in December 2006. The meeting took place in June 2007.
4) Dr. Nagarjun mentioned that someone who has faithfully implemented the specification still cannot read older documents, and therefore cannot reproduce them. Thus, only Microsoft can achieve the design goal of the specification. To his credit, this shows that he knew the primary design goal, that of creating a format that can be on par with older formats to represent richly formatted documents in storage. Now, I have two objections to his statement. One, how can he assume that the design goal means "faithfully replicate Microsoft Office documents", and not "faithfully and without loss of fidelity, store (a)documents that have pagination, applied styles, framed images etc. (b) spreadsheets that have features such as formulas,embedded charts etc. and (c) Presentation files with transitions, animation etc."? Two, is he saying that only Microsoft-created applications can open and edit the Microsoft Office binary file formats? Hasn't that problem already been solved to a great degree of success in his Open Source world? Isn't the availability of a format that by design allows storage of everything that they have succeeded in reading a good thing?
5) Several members persisted in their demand that all parties should also have access to the old Microsoft Office document standards. I do not necessarily object to this demand - I object to it being raised as a point to question the qualifications of the OOXML standard. Still, Microsoft seems to have listened, as of February 15, 2008.
6) Dr. Bhattacharya and Dr. Nagarjun seem to be somewhat confused about the use of custom XML inside OOXML documents. Their understanding, from what I can read, is slightly flawed. In any XML schema, there are provisions for extension. This is legitimate XML. In OOXML, such extensions may be used to add semantic meaning to existing document constructs such as paragraphs or tables. For example, custom XML allows for adding attributes to a table such that with a simple XML tranformation, the data in that table could be *also* produced as an XML infoset conforming to a customer-supplied schema. The key word in the preceding statement is *also*. This feature does not affect the regular purpose of the format, that of representing a richly formatted document, in ANY way whatsoever. And the contention by Dr. Nagarjun that this makes interoperability impossible is..well...unfortunate, and unfortunately typical.
7) The absent professor S.K. Gupta states that "voting for OOXML by BIS at ISO will increase the digital divide in India and will harm the domestic industry and economy." Nowhere in the rest of the e-mail quoted does he qualify the "digital divide" comment. He does go on to qualify the "domestic industry" comment though - "Moreover, the stated objective of OOXML is to support Microsoft office. Further, OOXML has only one known implementation . There is no need to make a standard out of a format that has only one known implementation. OOXML will clearly prevent Indian software firms from coming out with their own office productivity solution that is so critical for reducing the digital divide. Given the size of the OOXML documentation, it is also not practically implementable." I will not debate the so called "stated goal", and will not contend the "there is no need" statement. The good professor speaks his mind - fine. But I think he is insulting the intelligence of Indian software vendors and professionals when he says that OOXML will somehow "prevent" us from creating our own productivity solutions. How can the existence of a *storage format standard* prevent anyone from doing anything? Furthermore, creating an office productivity solution is "so critical for reducing the digital divide"? How, sir? Lower pricing for office productivity suites? Localized Office productivity suites? Aren't both options available today, when there are multiple (admittedly de facto) standards available?

There are three things which bother me in all this. They shouldn't, because all three are the way the world turns, and I should have got used to them by now. But they do.

From the minutes, the feeling I got was that the Microsoft representatives were on the defensive throughout. The Open Source proponent, typically, was vocal, caustic even. I call this the पिंजरे में बन्ध शेर को बच्चे भी मुंगफली फेंककर मारते हैं syndrome. After all these years, this still bothers me.

The supporters of anything other than Microsoft can use whatever language they feel like, say whatever they want, with no supporting evidence, or pointing to each other's work as evidence. And this is, somehow, okay. This bothers me.

The learned Professor is so...professor-like. Correction, current professor-like. This bothers me at a more personal level, and I will blog about this separately.

I applaud those that sat through this meeting. Some of the comments sounded vitriolic. I would not have been able to tolerate such. Incivility disgusts me.

* Guruji is Sanjay Vyas, who has taught several generations of Indian software professionals to think in Visuals and dream in C++.

** Amol is Amol Vaidya, one of the best .NET (and associated subjects) trainers in India.

I needed to quickly build a SharePoint tool for a friend, and all that I had access to was a Windows Server 2003 machine with WSS 3.0 installed. No dev tools, no SDK, and (for some very valid reasons) no way of downloading them either.

So I built the tool - 1 ASPX page, 2 compiled Web Server controls, 1 Feature, 1 Solution Package - using just a text editor (NotePad++, which happened to be installed. Very nice. I normally use Sciite.), the compilers included in the .NET Framework install, and the tried and tested method of copy/paste from existing feature definitions. I used a horrible method for generating guids involving SQL Server, uuidgen not being present.

The experience makes me feel good, good and bad. Good, because it's possible. Good, because so many years away from my "hardcore" days, I can still do things this way. And BAD, because I actually feel good about things like this. Looks like I still don't have a life.

Forms Authentication in SharePoint Products and Technologies (Part I, Part II and Part III) should be required reading for anyone studying SharePoint development. For the lazy progeny of bachelors, here are some salient points.

1. SharePoint can use ASP.NET Forms Authentication to authenticate users, and manage groups. This means that we can use potentially any security system to access SharePoint sites.
2. The same site can be made available to Windows and non-windows users, but through different URLs or Zones. Through careful configuration, permissions can be given to both kinds of users while accessing the site through either zone.
3. Using a non-windows authentication system affects the following features:
   • Accessing regular SharePoint sites
   • Using the "People Picker" user interface element
   • Using the "presence" features in SharePoint
   • Accessing and giving permissions in SharePoint Central Administration
   • Creating and using "My Site" sites
4. Creating a custom provider is possible using the ASP.NET 2.0 Membership system. This is the most complete way of integrating an existing security system with SharePoint.
5. Non-windows authentication systems can also be used to access SharePoint web services.
6. Non-windows authentication systems affect the settings of the two Search services in SharePoint, as these services crawl SharePoint content as a designated SharePoint user. SharePoint was designed to use a Windows user in this scenario, and although SP1 has allowed for using a non-windows user for this, one should still use Windows-based authentication for running the search.
7. Client Integration Features (opening and saving documents directly to and from SharePoint, for example) can be affected by using non-windows authentication.
8. The MOSS 2007 Single-Sign-On feature can be affected by non-windows authetication.

Details for points 1, 2 and 3 are in Part I. Points 4 and 5 are covered in Part II . Part III covers 6, 7 and 8.

Some time back, I was about to write a post that would reveal my feelings about the current state of Microsoft documentation to the two and a half people who read my blog, and simultaneously teach them some basic Hindi profanity. But then my blood pressure returned to normal. The Hindi profanity-laden draft sits in my space, never to be published.

Then today, via the Visual Basic Team Blog, I discovered and took the Visual Studio Content Survey. Please, two and a half people, do so too. Thanks for helping improve Microsoft's documentation!

I want to talk about some of my responses. The survey is anonymous, so I will not quote the questions themselves. But on the off-chance that someone from the relevant team at Microsoft is reading this (and cares), I will mention the question numbers. I have no problem in providing a signed response to the survey.

My answers to the survey are quoted in this color. They are accompanied by some additional comments I want to make.

Partial answer Question 12: Once upon a time, there used to be separate User Guides and Reference Guides for technical topics. How-to topics and samples would go to the User Guides, while Reference guides would contain both very well-written concept material, as well as very well cross-indexed references. I would like to see all Microsoft documentation go back to this format. I could then skim (and then ignore) the "user guide" section, [print out and] read the concepts part of the "reference guide" section page-to-page, and consult the reference part at need.

One thing I miss is properly written concept material. The last (chronologically) place where I saw acceptably good concepts documentation was the .NET Framework SDK 1.1. After that, we have been inundated with samples, "How-to" articles (please) and "Quickstarts".  These things have their place (yeah, in the freaking open-sore world. Okay, kidding. Yes, a lot of people benefit from the How-To and Quickstart articles (चू...ने हुए लोग).  Yes, it's Open SOURCE. Sorry), but so does the kind of writing I am talking about, which seems to have disappeared. Case in point - A topic called "Control Execution Lifecycle" in the .NET Framework SDK 1.1. The ms-help URL on my machine is ms-help://MS.NETFrameworkSDKv1.1/cpguidenf/html/cpconcontrolexecutionlifecycle.htm. Find me the equivalent topic in the .NET Framework 2.0 SDK, or the Windows SDK for .NET Framework 3.0. And look at the difference. In finding the damn thing in the first place, and if you do find it, in the content.

Partial answer to Question 13: It seems that the task of explaining concepts has been farmed out to blog writers, end user (MSDN "Wiki"? I mean, come on...) or worst of all, book writers.

This question talked about "content types". Apart from the usual suspects, it mentioned blogs, the MSDN "Wiki" and MS Press books. I don't like this. I do NOT want to reference any number of blogs to get accurate information on a released product. I do NOT want to write it myself, or read things contributed by other users which, and this is the clincher, may or may not be correct (In any case, user-contributed content does not go through a formal check. And no, "peer-review" does not cut it. Too many iterations, and too much time, and very real danger of something that "seems to" work getting accepted as gospel truth). And I do NOT want to pay extra to learn the basics about the software that I have just bought.

Blogs, books, and even "technical articles" inject a certain amount of personal predisposition...okay, strong term, let's make that "opinion"...into the topic being covered. That's a good thing; I buy Charles Petzold, read Raymond Chen, and admire Bruce McKinney. But I would like to read these opinions after, or at least during the process of, forming my own. To form my own opinion, I need an impersonal, just-the-facts, readable text OVER AND ABOVE a hyperlinked reference. This is what the old "Technical Reference Manuals" used to provide, and this is what is sorely lacking in these fill-in-the-blanks-in-the-Sandcastle-generated-outline-and-write-some-examples-and-माँ-का-आशिर्वाद-howtos-to-supplement days.

That's another thing: the Table Of Contents.

Different people browse content in different ways, which is why there are Contents, Index, Search and Favourites...sorry, Favorites options. For me, the Table Of Contents is the best way to locate and recall information. Even in paper books, I use the table of contents a lot. In current editions of Microsoft documentation, the Table of Content has gone smelly. It's badly organized, inconsistent, and worst of all, incomplete. Show of hands, those who could not "Sync TOC" from some MSDN topic or another. Can I see two and a half hands?

There was a question about the content being available exclusive online - as in - via World Wide Web.

NOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO!!!!!!!!!!!!!!!!!!!!!!!!!

What do I do on the road? At 11pm on a rainy night when my ISP's router is under water (I live in Mumbai, and don't have a life. MSDN is after-dinner reading.)? And when I...let's just say that MSDN makes a nice alternative to a newspaper or magazine. So please, make me pay for the DVD burning and shipping maybe. But not online, exclusively. And my favourite method of browsing the content, the TOC, SUCKS on the web.

Question 13 also talked about videos etc. as content types. You know, I can read text MUCH faster than I can watch a video. And text content can be much more distilled than video or audio. So, I would not like to see any content that is video or audio exclusively.

There was the usual "Anything else" question also. My answer:

Please, make sure the documentation is available completely offline for those who require it. Please, make sure *every* topic syncs to the TOC. And please, either give up on "integrating help collections" or get it right.

Preferably give up, on that last one.

In summary, documentation needs some real writing, by some real writers. I'd like to see the day when the FIRST (as in, the first place an end user looks) authoritative source of information about a released Microsoft product is NOT a blog from the product team, NOT online MSDN, NOT a technical article, NOT a public standard, and definitely NOT a book by a distinguished author, but the documentation that accompanies a product.

Maybe we should go back to printed manuals. For some reason, I feel people take more effort if content has to be in print.

क्या लं...बा हो गया। पोस्ट।

I received a large zip file with many levels of folders within it. I unzipped this inside a folder that was itself five levels deep in my folder hierarchy, using the built-in zip functionality. Once the full path names exceeded a certain length (~200 characters, I think. I’ll update this post after some experiments later), the unzipped filenames were *silently* truncated.

I searched for reports of this behavior, but couldn’t find anything. The truncating part doesn’t bother me much; the silent part does.

What does the following code do?

It's a utility that I wrote for myself. Lots of comments have been removed, and a couple of names have been changed to protect their identity. But it shouldn't be very difficult to figure out what this does.

@echo off
GOTO :Main

REM "Subroutine" [snip]. Call with :GetRegValue  
:GetRegValue
REM Careful here. The value after delims= is a TAB
FOR /F "tokens=1,2,* delims=	" %%i IN ('REG QUERY "HKCU\Software\Microsoft\Windows\CurrentVersion\Internet Settings" /v %1 ^| FINDSTR "%1"') Do SET %2=%%k
GOTO :EOF

REM "SubRoutine" [snip]. Call with :NameChanged1 
:NameChanged1
SETLOCAL
SET MyProxyEnable=""
SET MyProxyServer=""
SET MyProxyOverRide=""
CALL :GetRegValue ProxyEnable MyProxyEnable
CALL :GetRegValue ProxyServer MyProxyServer
CALL :GetRegValue ProxyOverRide MyProxyOverRide
echo ProxyEnable$REG_DWORD$%MyProxyEnable% >%1\ProxySettings.RGA
echo ProxyServer$REG_SZ$%MyProxyServer%