What you see everyday...

re: What you see everyday...

anturio — Tue, 08 Sep 2009 10:44:17 GMT

Happy Labor Day Weekend!

In case I forgot to mention

einhverfr — Wed, 28 Jun 2006 06:01:12 GMT

My two favorite RDBMS's are both open source: PostgreSQL and Firebird.

re: What you see everyday...

einhverfr — Wed, 28 Jun 2006 05:50:00 GMT

Cronos;

I will confess that MySQL is my least favorite RDBMS (somewhere just behind MS Access), so I am a little biased here. Most of that bias is based on the frustration I usually experience when I have to work with MySQL....

The real problem is that MySQL doesn't just "extend" the standard. It *breaks* the standard in numerous ways even when run in ANSI mode. And since most applications are not built on systems running both strict and ansi mode, so you have to deal with these major issues when porting a non-trivial application. In short MySQL employs the same "partially embrace, extend, extinguish" tactics that people accuse Microsoft of pursuing.

My favorite RDBMS has its share of extensions but most of those are sufficiently wrapped in the database layer as to make porting at least possible without unduly modifying the application. But porting a large app is never simple...

Best Wishes,
Chris Travers
Metatron Technology Consulting

re: What you see everyday...

cronos — Fri, 23 Jun 2006 18:00:43 GMT

einhverfr: the SQL Server also has extentions and non-standard features:

http://www.tdan.com/i023fe03.htm

The differences with free software options are:

the SQL Server and other propietary options costs a lot, and the free software ones don't;
and if you really need that features, and you have good programmers, you can modify the free software, but you can't modify the propetary software.

Only for the cost and features of free software databases I'm convinced there are the best option, but if you appreciate the freedom too, there is noting to do with propietary software.

re: What you see everyday...

rjdohnert — Fri, 23 Jun 2006 06:46:20 GMT

Great post bill. Very informative and interesting. keep up the good work.

Migrations and vendor lockin

einhverfr — Fri, 23 Jun 2006 04:23:35 GMT

I am fairly familiar with the issues of porting MySQL to PostgreSQL. I have ported a few applications, maintain an application that used to support both, and have even authored a simple guide which is available on my whitepapers page.

The problem is that MySQL has a large number of "features" which either "extend" or outright violate the SQL standards in ways that are usually incompatible with other systems. Case sensitive table names, non-standard operators, lax error checking, and so forth mean that if the application has never supported other RDBMS's in the past, porting is likely to be a bit of a task.

Best Wishes,
Chris Travers
Metatron Technology Consulting

re: What you see everyday...

cronos — Thu, 22 Jun 2006 18:25:32 GMT

einhverfr: I done it sometimes, from MySQL to PostgreSQL, and from PostgreSQL to SQLite. What's the problem?

Is more easy migrate from Access or SQL Server to another database? I searched this at Google, and for Access and SQL Server, I didn't find anything free. I found at Microsoft a software for free to migrate oracle to SQL server, only after filling a lovely web form:

http://www.google.ca/search?hl=en&q=migration+access&btnG=Google+Search&meta=
http://www.google.ca/search?hl=en&q=migration+SQL+server&btnG=Google+Search&meta=

And, by the way, lots of free software and howtos to migrate from and to MySQL and Postgres, and even to migrate from Access and SQL Server:

http://www.google.ca/search?hl=en&q=migration+MySQL+Postgres&btnG=Google+Search&meta=

Talking about vendor lock-in?

Reply to Fluke

einhverfr — Thu, 22 Jun 2006 04:12:06 GMT

Fluke wrote:
'Which brings me to the next point, the biggest item that stood out for me in the translation I read was "will act with intention to search the reduction of the dependence of only suppliers." Computer Associates and Microsoft have become known for creating methods of vendor lock-in. It tends to be far easier to transition between one FOSS vendor to another while retaining the same level of functionality than it is to transition between vendors that promote prioritary feature/extentions. To steal a line from the popular press, "Microsoft is known for it's 'embrass and extend' adoption of standards." '

You have never tried to migrate away from MySQL have you?

Well, I guess that is a bad example because for all its source license under the GPL is very much a proprietary product.

Best Wishes,
Chris Travers
Metatron Technology Consulting

re: What you see everyday...

fluke — Wed, 21 Jun 2006 16:48:02 GMT

You seem to be leveraging the myth that Microsoft is only a "software" company. For purposes of *budgets*, they tend to be equipment.

There are several companies that consider the cost of OEM pre-installed software to have come from the equipment budget. That means those companies don't even factor in the cost of Windows OS and Office as part of the software budget. Is Brazil that much different?

If the Dept. of Health assist in the purchase of a MRI or CAT, does none of that money go to Microsoft? Does it count as purchase of software?

Of the R$2.2 million that Brazil spends on cellular telephone, does none of that go to Microsoft? Does use of Window CE on a cell phone count as purchase of software?

Another hidden cost to any country is how choices made by government for running software gets passed on to it's citizens. Adoption of the Apple II+ was successfully largely due to people buying what their children had already tried in the schools. Such leveraging of goverment institions to gain acceptence continues. In the US, each time April 15th rolles around, Adobe gains several more users of Acrobat Reader since the IRS seems to favor PDF. But if anyone wants to leverage this install-base, they have to accept that PDF is not a truely free standard (as shown by Adobe's reaction to Office 2007). Governments are the best place to make or break general acceptance of free or prioritary standards.

Which brings me to the next point, the biggest item that stood out for me in the translation I read was "will act with intention to search the reduction of the dependence of only suppliers." Computer Associates and Microsoft have become known for creating methods of vendor lock-in. It tends to be far easier to transition between one FOSS vendor to another while retaining the same level of functionality than it is to transition between vendors that promote prioritary feature/extentions. To steal a line from the popular press, "Microsoft is known for it's 'embrass and extend' adoption of standards."

To avoid Microsoft as an example for a moment (some people seem to take it as being anti-MS to talk about the company at all), Kodak was known for a while of providing discounted cameras that took prioritary batteries. The initial cost for the initial set of waves of costs might be low, but the vendor lock-in for batteries end up increasing the TCO. If the vendor decides to stop providing the batteries then the cost may become higher as a new initial investment in a camera must be made.

Microsoft "cameras" tends to come with non-standard batteries. A standard Kerberos ticket is normally only provides information for authentication. MS-AD tickets use a reserved field to accomplish embrass and extend such that authorization also becomes part of the ticket (something normally best handled in LDAP). The MS-AD ticket server can not be swapped for a third party Kerberos/LDAP servers and still expect to get the same level of functionality even if both servers are fully compliant with all the provisions of the RFCs. This lock-in methodology is impossible to enforce when the customer is using strictly FOSS.

Even more upsetting to citizens of smaller countries that can't keep up with the US life cycle for software is the lack of control on the road map for the software. Microsoft may have had a road map for Windows 98/ME to include security fixes til June 30th, but MS06-015 clearly shows that the practical road map for security fixes ends earlier. In the FOSS world, when Red Hat decides to terminate support faster than the customers are ready to accept then Progeny can step in and provide the patches without any authorization from Red Hat to do so. There will be no "Progeny" patches for Windows 98 or ME as access to the source code to either OS does require authorization from MicroSoft. The customer is lock-in to a single source for "batteries" (or in this case patches) and without them to accomplish the same task of continuing to get security updates requires a premature upgrade of the OS (and in several cases the hardware to meet the minimal requirements for the upgrade).

Reply to Dreamer

einhverfr — Wed, 21 Jun 2006 15:27:14 GMT

First, Microsoft is doing a number of source code releases both as "Shared source" and open source. And they are working with some open source projects in some ways too.

However, if you are suggesting that the source code for Windows or Office needs to be released under an open source license, that is anything but simple. Usually it takes an army of lawyers to review licenses of existing technologies that Microsoft may have with other firms and a massive engineering effort to replace those parts that the business does not have a right to release under such a license. A year on and OpenSolaris is not yet even complete. So this is a major step in terms of effort and a huge commitment of money. And it wouldn't really help Microsoft in any way I can find at the moment-- they business is not ready to be an open source business.

I have a unique perspective because I used to work at Microsoft and was even involved in discussions about how to compete with Linux. These days I run my own open source based business.

BTW, I am sure that Mr. Hilf is not allowed to tell a shareholder what he/she should be doing with MSFT stock at the moment by SEC rules. At least when he is acting as an agent of Microsoft.

But Microsoft has a huge problem and I think you are largely missing it. Microsoft simply cannot expect the kind of market growth that they have experienced over the last few decades. Their growth is slowing in large part due to saturation of the US and EU markets and an inability to sell their product effectively in the rest of the world. With hardware remaining in service longer, software upgrades are a harder sell. Why else do you think they want to offer Software Assurance?

Where does Linux and open source fit into this picture? First, these products provide genuine competition. Right now, in many areas, they are the *only* viable competition left. And competition provides 2 things:

1: It provides a real alternative infrastructure for a business network. And the ideal Linux/UNIX structure (seamless, shared everything) is very different from the ideal Windows structure (loosely coupled).

2: It provides market leverage against Microsoft. Thus Microoft must often offer steep discounts to keep major players on their platform. It also makes the success of offerings like Software Assurance less likely.

re: What you see everyday...

carlos toguchi — Wed, 21 Jun 2006 15:16:01 GMT

The cost of Microsoft Software is big, very big, and the cost of OEM licensing ???
You it entered this cost?

re: What you see everyday...

dreamer — Wed, 21 Jun 2006 12:30:41 GMT

"Linux-killer case studies are not working, let's put a OSS lab in place and do some massive brainwashing."

I think so.

If Microsoft really wanted to get into open source, they'd release source...erm, wouldn't they? It's quite simple.

Bill, the reason that Microsoft stock has fallen rapidly (like from $28 to $22 in a matter of weeks) is the continued disregard for Open Standards, lack of sourcecode provision, EU antitrust cases and dangerous EEE (Extend Embrace and Extinguish) tactics.

It's not about money - it's about freedom. You are running into some serious issues with software licensing and your EULA. 'Freedom of information' is key now. Proprietary file formats only lock up the data and this is the biggest fear at acquisition. What if? People are scared.

My clients would pay several times the cost of Microsoft licensing for alternative solutions: 1) Because the solution is 'free' from vendor lock-in and 2) Because all of their money can go towards support, not licensing.

The public / business software market has changed - it's solutions as services. That's just not Microsoft's model. So, it's either change time or quit time.

The problem for you is that time is running out. Shareholders want the next big profit - and that's not going to come from Microsoft.

There was a unique opportunity in the past few years to embrace Linux and OSS. OSS was a movement, not a company to try to wipeout. When you tried to shut it down: A) You couldn't - it wasn't a company and B) All you did was offend developers - and developers are the channel to desktop users.

The unfortunate fact is that OSS is actually more cost-effective, by 20-50%, with a greater level of suppport and security (enforced user not admin etc..). And it's free (as in beer) to learn - so everyone who wants to can get experience (the next big economies). So, really you're in trouble.

The options - either you become a console entertainment vendor (didn't work) or you ship games for your platform (PC market sideways in favour of consoles) or, erm. - give developers what they want...Linux.

Mr Gates was quite right when he said "No-one owns the factory for ideas.." I'm just confused then, why you continue to try to close the factory down?

Microsoft needs software patents in Europe - to make more profit from the $16bn a year that goes through Ireland and avoids EU taxes - for example in the UK and at the same time avoids US taxes. Unfortunately, that vote was lost by MEP's in a 648-14 vote against.

So, no more savings, no more profits. You see, when the innovators create - they don't say 'How much can I get for this?' they say 'How much can I give with this?'. The money comes - very quickly - offers from all sides from companies that want to make a profit - but it's on the terms of the developer. The innovators aren't inside Microsoft and unfortunately the MSFT shareholders have just found out.

I don't wish Microsoft financial difficulty, but it's looming. Maybe you could pull back something with a Linux distro, or with inclusion of Open Office or something - but you know that you're doomed once you try to make the bridge there. People will say 'Hey, why not just use Linux with these applications?'.

You know what I see everyday? People - people with growing IT skills. People who are more and more abstract from the Operating System. People with mobile devices. People using Web Services. And these people want all the file formats in the world. The next generation only want the 'freedom' ones.

That's the only option. Still, it might be too late for the stock value. I'm just interested - what's the tipping point for Microsoft? 40% off the share value... 60%? Or less? 20%? Perhaps we're already close...even on the wrong side of it?

MOST IMPORTANTLY: Can you tell the shareholder why he/she shouldn't sell MSFT _right_ now?

re: What you see everyday...

mleite — Wed, 21 Jun 2006 08:04:10 GMT

First, I see "Microsoft is just barely on the top spend list". Then, as a P.S. "they enter into their 2nd, 3rd, or 4th ‘wave’.. and realize there are still costs with free software".

So, I think I got the message: "Microsoft is not on the top list but we desperately need to be". Linux-killer case studies are not working, let's put a OSS lab in place and do some massive brainwashing.

re: What you see everyday...

einhverfr — Wed, 21 Jun 2006 06:06:48 GMT

Hi Desert_dev;

I can sympathize. I was doing a security review of an application we provide a fair bit of suppot and I discovered that the approach to authentication was fundamentally flawed and trivially circumvented. After a series of arguments with the maintainer of the software, I have concluded that I will need to fix it before I make a security announcement because if I don't it won't ever get fixed. Fortunately I can fix it.

Shows that open source by itself is no guarantee for security, but in the end, maybe it helps.

Best Wishes,
Chris Travers
Metatron Technology Consulting

re: What you see everyday...

desert_dev — Wed, 21 Jun 2006 04:49:56 GMT

Chris:

> Services and support are a far wider field than just breakfixes.
> Indeed, I would argue that customization and engineering are a major
> part of that. Indeed the idea of support is to support or work with
> the IT operations of a business to solve whatever problems they can't
> effectively tackle.

Definitely agreed - particularly in embedded. Consider this article -
it states that embedded is purely a service industry, and shouldn't
be product based at all:

http://linuxdevices.com/articles/AT6989225710.html

At the IT level, I see customization as a very different thing -
more customization of existing applications (like interfaces for
a site on top of a content management app) than specific development
for a client. That's probably because embedded is a much less mature
industry in a lot of ways, but it definitely is a large piece of
the pie. Obviously IBM has some leverage here.

> There are a wide range of such services. And you have a point that
> there is a lot of shoddy FOSS engineering out there. But there is
> also a lot of high quality FOSS engineering and a lot of shoddy COTS
> engineering.

Couldn't agree more. :) As long as the industry has engineers, there
will always be good, and always be bad. Some write COTS, some write
FOSS. I seem to be a dowsing rod for bad FOSS software that has to be
made to work.

The trouble with FOSS is that when it's bad, you get to see the shaky footing. Consider a closed in COTS bridge - at least you can have the
(potentially false) sense that it's well built. With a FOSS bridge, you may look down and see the whole thing holding together by a string.

Which is better? It depends if you can fix the bridge yourself, or if
you just want someone to go after when it all comes crashing down.

I favor the open bridge myself, and I try not to look down except
when I have to.