Public Class GeoffAppleby : Will the MS Dev Team Ever Learn?

Navigation

Archives

Will the MS Dev Team Ever Learn?

Me, I have no doubt :) But lately I'm seeing more and more comments from people (which is the right of every blogger out there, mind you - speaking their mind, that is) that seem to think that they won't.

What's sparked me off this time? Scott Bellware posted recently about his idea for Adopt-a-Softie (BTW dude, you better get yourself a trailing ^(tm) on that term before someone else steals it!), with the idea that some of the dev team should join the international student exchange program, spending time working on the average guys development team for a month.

Don't get me wrong, I'm all for everyone having a voice - and your voice has a right to be heard! But in the same breath, it of course means that I get to have a voice too :) I have to admit that lately it looks like all I'm doing is doing my darnedest to pick on the guys I left behind at CodeBetter, but it's certainly not the case - they're a fabulous and fabulously smart bunch of guys, but as a result they get me thinking the most. There's certainly no offence intended :)

So I was going to just add a comment in response to Scott's post, but as I was typing it, I decided I might make it a bit more public than just a comment. Some things are too thought provoking to be just another comment feed.

So to Scott, I have basically two things to talk about. The first is the concept of your Adopt-a-Softie. Do you realise how close you are to what they're actually doing at the moment as we speak? You thought you were (or it seemed that way at least) proposing something that wouldn't actually be listened to, but did you see one of Soma's most recent posts? It sounds me like they're having the same sort of idea.

It's not an exact match, and 3 days is certainly a whole lot less than a month, but it's going in the right direction isn't it?

I'm not saying that you've got nothing to complain about, Scott. It's just that lately I'm seeing a lot of negative thoughts directed at the developer platform (not you in particular, it's around a lot of places), but no ones seems to be paying attention to the fact that they are actually trying. The thing about such a large corporation is that it takes a long time to get the wheels turning. Me, I'm seeing the glass half full, and at least taking hope in the fact that, however slowly, things are starting to get there.

The second thing is more of a question. You said:

Really, I'd prefer to use a Microsoft-built tool, but I loose progressively more faith in Microsoft developer tooling as it continues to trade credibility and quality for lesser goals.

The whinging (that I've seen from everyone) has been about VS2005. Now, I might be wrong, but it hasn't actually changed since it RTM'd last year has it? The code hasn't morphed into something else? What exactly is happening to make you think that they've been trading credibility and quality?

I'm a bit disappointed in VS2005 too. There's a lot of cool stuff, and there's stuff that seems to have taken a step backward. but there isn't that much stuff that's gone back. I think so anyway.

The biggest sore point for me was the major screw up in removing project files from web projects. But they listened, and they've brought them back. I don't know who was filling in the responses in the Product Feedback Centre, but back before VS came out it really did feel like they weren't listening - but once they had time to breathe, we've discovered that they are listening, and trying to make up for it.

So how can things have been getting steadily worse over the last six months?

I'm holding out only a small amount of hope for Orcas. Maybe there's less communication/openness in the blogs lately, but I'm not seeing much by way of any major changes to VS in the next version - things seem to concentrating solely on getting LINQ going, and the WinFX bits, and getting a version that works well on Vista when not running as an admin sorted out. That's probably a lot of work, so I wouldn't be surprised if there's nothing else there.

But the version after that? Man, I've got my hopes up for that one :) But I think (so far) that I won't be too terribly disappointed. For me, I'm still seeing signs that they're all trying, and trying hard.

Posted: Tuesday, 2 May 2006 4:01 AM by Geoff Appleby

Filed under: Rants, Geek

Comments

Scott Bellware said:

> Now, I might be wrong, but it hasn't actually
> changed since it RTM'd last year has it? The
> code hasn't morphed into something else? What
> exactly is happening to make you think that
> they've been trading credibility and quality?

Over the past six months, I have hit more and more function points through more use of the tool in more diverse applications. So it would be more accureate to say that my perception of the tool is changing rather to chance the implication that the tool iteself is changing.

To wit, my declining comfort level with Microsoft product creidibility isn;t limited to Visual Studio. In the past little while, a lessening of usabibility seems to be expressing itself in other Microsoft products that I use as well.

I agree that Microsoft has done truly great things with the developer tools lately, but for me, they are offset by the unessesarilly frustrating things they've purposfully built into the tools out of what I suspect is a lack of understanding of their customers at large. It's hard to remain postive with Microsoft's moves when they often amount to one step forward, two plus or minus one and a half steps backward.

I hadn't seen Soma's post. It's an interesting idea, but I seriously doubt that 3 days amounts to much impact. I also suspect that Microsoft chooses the customers it interacts with in this program with based on whether the customers are either the biggest accounts, or whether the customer's approaches are already reasonably in-line with Microsofts existing pre-conceptions about tool use in the wild.

I don't believe that Microsoft could change on a dime if it wanted to. The size of the org certainly plays a role, but its generalized lack of agile practices is a more significant contributor.

# May 2, 2006 11:37 AM

Dave Burke said:

Always good to see a fresh rant from the Geoffster!

# May 2, 2006 2:03 PM

Tobin Titus said:

Scott continually beats all the worlds problems with the agile hammer because all the problems he sees are nails. It shows an incredibly narrow view of how the world develops software, yet Microsoft is the one to blame simply because they don't develop software the way that Scott develops his software. Honestly, Microsoft wants to please the Scott Bellwares of the world -- but it doesn't expect to. As long as squeeky wheels get the grease, you will always have Scott Bellwares clamoring for attention at the most base levels. All I can say is , to date, he hasn't provided a single solution to any one of his misdirected rants. That should speak volumns by itself -- no additional blog post required.

# May 2, 2006 2:53 PM

Geoff Appleby said:

Hey Scott,

>Over the past six months, I have
>hit more and more function points
>through more use of the tool in
>more diverse applications.

And are you planning on blogging these specifics? No one can do anything to fix them if you don't talk about the actual problems you encounter on a day to day basis.

If you discuss the problems, people at MS might (dare I say, WILL) see the report, and do something about it.

On top of that, other people might come up with workarounds that you could use while waiting for a 'fix'.

>the unessesarilly frustrating things
>they've purposfully built into the
>tools out of what I suspect is a lack
>of understanding of their customers
>at large

This is where I disagree. I don't think it's lack of understanding of the customers at large.

My feeling is twofold: 1) They've covered 90% of cases for 90% of the people - and if they did anything different, they'd have a lot more people complaining and 2) Time constraints - it takes time to get stuff right, and i think there's a few things that could have been done better but they couldn't spend any more time on.

Some things don't ship until things are _right_, other things ship as 'good enough for now, we'll improve it in the next version'.

I'm fine with both cases, so long as they _do_ improve it in teh next version.

>I also suspect that Microsoft chooses
>the customers it interacts with in this
>program with based on whether the
>customers are either the biggest
>accounts, or whether the customer's
>approaches are already reasonably
>in-line with Microsofts existing
>pre-conceptions about tool use
>in the wild

You assume too much :)
Well, i don't know if you do, but do you know? You're basing all this on assumptions. Why don't you go and find out first?

>but its generalized lack of agile
>practices is a more significant
>contributor

And again we come back to that. *sigh* Why does it? Is this some magic bullet? Is it the only bullet?

Why the hell does 'agile = right' and 'everything else in the damn world = wrong'?

I'm so fucking sick of these statements. Agile is ONE option out of many that work. It's not the only one.

# May 2, 2006 6:22 PM

Jens Samson said:

> Some things don't ship until things are
> _right_, other things ship as 'good enough
> for now, we'll improve it in the next
> version'.

> I'm fine with both cases, so long as they
> _do_ improve it in teh next version.

The real problem is when things get inworse instead of getting better. The VB.NET designer is a good example of that. I have a VB.NET 2003 solution where components and controls are in one project file, the composed components and controls are in another, and all these are used in another WinForms project. This used to work, but in VB.NET 2005 it does not anymore. Sure, it works when you fire up the IDE, but as soon as you start changing code or start recompiling, the designer gives up and ony displays Exception messages. I'm not the only one who has this problem. It is so problematic that I've reverted to VB.NET 2003, despite missing Generics, Edit And Continue and Refactoring.
I can't understand that someone thought that the designer is 'good enough for now', it isn't.
My overal feeling with VS.NET 2005 is that they tried to do too much at once. Out of experience I know that that usually leeds to unstable software.

# May 2, 2006 11:15 PM

Scott Bellware said:

> Why the hell does 'agile = right' and
> 'everything else in the damn world = wrong'?

> I'm so fucking sick of these statements. Agile
> is ONE option out of many that work. It's not
> the only one.

It's one option that works much better than most. You're still free to choose any option, just as you're free to chew glass rather than Juicyfruit.

A flat Earth is just one perspective of our corner of the cosmos, as is a Sun that revolves around the Earth. And yet, these both turned out to be wrong, and some of the people who advocated contemporary models were executed by the keepers of the traditional faith.

How certain are you that agile development isn't the paradigm shift that its practitioners rave about?

What does your hands-on experience with agile practice tell you?

If you were a natural philosopher in Galileo's time, before you would be able to validate for yourself that the Earth is indeed round and that the Sun is at the center of our planetary system, you'd need to become intimate with a lot more hard science than what was required of you in maintaining a 14th century natural philosopher's status quo.

In your experience, what are the principle design expectations of practicing contemporary Test-Driven Development in static environments? How are they achieved?

It sounds hokie to say, but once you can begin to answer these questions for yourself, you'll likely be on well down the path to delivering better software that you might have expected of yourself previously, and you'll also know a lot more about why agile isn't like traditional development and therefore why it's not on par with traditional development.

Again, and again, and again... I've never seen a blog post, a series of blog posts, or a book that communicates agile even fractionally as well as practicing agile does.

The only thing in my life that I can compare it to is my meditation practice. I read about meditation for years. It never really compelled me sufficiently to give it a serious go. When I dabbled in it, I never really got anything from it. When I finally committed to an experiment to really give it a go for three months, it completely blew my mind and changed many fundamental ways that I perceive the world around me.

It doesn't come from reading. It's a practice. If you want to understand it, practice it. I guaren-freakin-tee you that after three months you'll be a different developer and you'll never go back to what you did before. And, the very best you could offer to someone asking you to communicate it to them would be to tell them to practice it.

Provided that I could find others to back the experiment, I'd bet you a month of your salary that if you spent only a month on my team that it would change you forever and that you'd have a very hard time trying to return to your old ways.

# May 3, 2006 10:36 PM

Scott Bellware said:

> Scott continually beats all the worlds problems
> with the agile hammer because all the problems
> he sees are nails

I beat all of my software problems with agile hammers because they consistently solve these problems better than anything else in my kit.

How about you? What was your experience of working on an agile project with experienced agile practitioners?

> It shows an incredibly narrow view of how the
> world develops software, yet Microsoft is the
> one to blame simply because they don't
> develop software the way that Scott develops
> his software

Microsoft is to blame for a lot of lack-luster software and depressed expectations for software, both its own and that of its customers, because it encourages the use of olden software development approaches through tools built by those approaches to support those approaches.

> As long as squeeky wheels get the grease, you
> will always have Scott Bellwares clamoring
> for attention at the most base levels. All I
> can say is , to date, he hasn't provided a
> single solution to any one of his misdirected
> rants. That should speak volumns by itself
> -- no additional blog post required.

Here are some of the base ways in which I have clamored for attention while not offering solutions:
http://www.netdug.com/Default.aspx?tabid=82
http://www.bellware.net/Events/TddSanDiego.aspx
http://www.bellware.net/Events/TddBayArea.aspx
http://www.bellware.net/Events/TddHouston.aspx
http://www.bellware.net/Events/TddMontreal.aspx
http://www.bellware.net/Events/TddDallas.aspx
http://www.bellware.net/Events/TddAustin.aspx
http://austin.innotechconference.com/Event/Austin_2005_Events/TestDriven_Development.php
http://www.adnug.org/codecamp2006/codecamp2006downloads.aspx
http://www.devteach.com/PreConference.asp#PreTDD
http://www.devteach.com/PostConference.asp#PreTDD
http://www.devteach.com/Session.asp

I don't actually believe that I can bring about a solution with Microsoft anymore. I exercise my prerogative to lament this state of affairs.

# May 3, 2006 11:03 PM