Peter Seale's weblog

Tuesday, September 23, 2008 3:10:01 AM UTC 

Summary first, because I just re-read this and it's ridiculous long

If you've subscribed to my blog exclusively for the SharePoint bits, sorry; now is the time to unsubscribe.

That goes for both of you/half my subscriber base :)

I've become increasingly frustrated with SharePoint (I'll explain why in, um, detailed fashion), and instead of souring like a lemon, I'm just going to refocus. This means I'm dropping out of the SharePoint blogosphere to focus on long-term fundamentals; knowledge that won't expire when SharePoint 14 hits beta. I may jump back on for SharePoint 14, but it all depends.

UPDATED 2008-10-18: fixed grammatical and factual errors. By factual errors, I mean that time when I attributed something to Chris O'Brien's blog, when oopsies! it wasn't him at all.

Introduction

I've become increasingly frustrated with SharePoint recently (again, I know), but what's been bugging me more lately is the fact that I'm solving all the wrong technical challenges. I think I have a (deeply suppressed) aesthetic sense, that every so often, rears its ugly head. Or flares out from my neck like a humongous goiter. Goiters are a serious problem—hey!—used iodized salt, it's that easy.

I get agitated at times staring at (undocumented) CAML, or anytime I work with InfoPath, or remembering when to surround hardcoded GUIDs with curly braces (and when not), or knowing when to dispose SharePoint-created objects, or remembering which Workflow activities only work in SharePoint Designer and not Visual Studio, or working around concurrency bugs, or troubleshooting failed Solution deployments. And so on, pick your topic; these are real issues by the way.

So what's the point of this post again?

I want to send the message out there to everyone working with SharePoint, to look beyond your immediate technical challenge, and think. Think beyond your immediate issue, beyond with what you're wrestling at this particular moment, beyond your immediate (sometimes overwhelming) technical challenge with me. Are you ready? Looking at the size of this post, maybe not!

Let's DO THIS.

Are you an expert at SharePoint development?

I recently watched Dave Thomas give a talk about developing expertise. It's interesting, go watch it if you have time; he's not making it all up on the spot, he relies on some kind of research. In the talk he does something I like: he categorizes everyone (using the Dreyfus model) somewhere along the expertise scale: Novice, Advanced Beginner, Competent, Proficient, Expert. And gives definitions for all five categories, and gives helpful tips for how to work with people of varying skills. It's not all fluff, go watch it.

Bringing this back to SharePoint: after trolling the entire SharePoint blogosphere for quite a while now, I've come to the conclusion that there are no expert SharePoint developers. I'd go as far as to say that there are very few proficient SharePoint developers, and you probably know all their names (hint: look for "MVP" in title and/or multiple Bentleys in the parking lot).

The rest of us are just struggling. The "everyone's struggling" theme really sunk in recently when I browsed the source code for SharePoint-tagged CodePlex projects. All of them (save one) disposed their objects improperly. Including mine, to be fair, I'm not hating on your free, labor of love project, I'm making a point. The point isn't to hate on your baby, the point is to say that no one is doing SharePoint development properly. Look on CodePlex; you'll see what I mean.

Let that sink in. No one is doing SharePoint development properly. And hey, your team may have tackled all the object disposal issues, but maybe instead, you're ignoring large portions of the framework and (unknowingly) rewriting large portions of it. Maybe you don't bother packing things in Features and Solutions; maybe you spend too much time packing things in Features and Solutions. Maybe you're great at writing Site Definition CAML but haven't gotten the memo that putting everything in your site definition is a bad practice (or the other memo telling you that custom Site Definitions don't upgrade to v14). Maybe you're unknowingly triggering framework bugs that ruin your customers' trust in your solutions. Maybe you write unmaintainable code.

Maybe you don't understand your customer's core problem!

Maybe, in your metaphorical dwarven greed, you've delved too deep into the framework, and unknowingly stirred the framework Balrog. And you don't want to wake the framework Balrog, let me tell you.

I don't know about you, but I'm there—I'm struggling.

Takeaway: just try to be competent

Dave Thomas says in his expertise talk that advanced beginners think they're much further along than they are, and those who are proficient or experts think they're beginners. Let me tell you: as far as SharePoint development skills go, I'm an advanced beginner, bordering on competent. And I mean it—knowing that he says "beginners tend to think they're experts, and experts think they're worse than they are"—with all that said and digested, I still think I'm a beginner (advanced beginner, I'm not a total scrub :) ).

So, and this applies to those of you who are still new to the SharePoint world and/or isolated—it's okay to admit that you're not an expert. If you need an ego boost, go look at CodePlex—you'll certainly learn a great many things from others' code, but you'll also realize that hey, these people don't have it together either!

Becoming a competent SharePoint developer

What I've also begun to realize, is that the more I learn about SharePoint, the less I want to apply it to different scenarios. Where before I used to say "SharePoint Workflow is supposed to be good at solving that problem," now I say "SharePoint Workflow is painful, and for long-term maintainability I'd recommend you manually code up whatever you need, elsewhere." Whereas I used to say "InfoPath is useful for simpler scenarios," now I further limit InfoPath to "end-user tool only; don't you dare use InfoPath's code-behind." I say these with specific scenarios in mind, I have biases, your mileage may vary, all things within reason, etc, but the point I'm making stands—I'm recommending SharePoint less.

Dave Thomas says, again, that as an advanced beginner you believe that nothing works. Unfortunately, every time I encounter a new piece of the SharePoint framework, I'm an advanced beginner. I don't have the opportunity of being the CQWP guy, every day. Today I'm that guy, tomorrow I'm the InfoPath guy, day after I'm the admin, day after I'm the help desk. I'm that guy.

And hey, I've been an advanced beginner before. For example, I distinctly remember the first time I tried to do Active Directory scripting with PowerShell. It was painful. But gritting my teeth and working through the pain, I became comfortable with it. And, now that I know my way around, I'm able to do some cool (simple) things with PowerShell and Active Directory.

This is a common pattern—I've been an advanced beginner elsewhere, worked through the pain, and by the end, I was comfortable and was feeling more confident.

What I'm trying to articulate in these long rambling paragraphs, is that as I learn more about SharePoint via experience, I become comfortable with pieces of the framework and trust the framework less. As I learn more, I trust SharePoint less. Today I'm totally comfortable in telling you that InfoPath is absolutely great for simple data entry, for simple review scenarios, but nothing else*. And this isn't just InfoPath. As I work my way around the framework, I am becoming comfortable with individual pieces, and as I do so, I am trusting each piece less.

And let me set the record straight. I'm less interested in whether a particular technology is Turing-complete—whether it can theoretically solve any problem—I'm more interested in whether using that technology is a good idea in the first place. And as I learn more about SharePoint, the world of good ideas seems to shrink.
*note for InfoPath nitpickers: yes, it's useful for other things, but yes, it's also easier for me to say "good for nothing else" than rattle off 15 edge cases.

Do you want to become proficient?

I think what is bugging me most is that I don't want to be a SharePoint apologizer. Not apologist, apologizer.

Oh, sorry, that doesn't work. Sorry, yeah, there's a bug. No, yeah, you heard it was great for ECM; well, yeah, it's cheaper, but we'll have to customize a lot. Oh, yeah, that's undocumented. Oh, no, we hit the list size limit early. Yeah, no, SharePoint Designer is unmaintainable, don't build solutions with it. Oh yeah, factor in cost of upgrades on every SharePoint project.

Last week we found out "the hard way" that SharePoint can't handle 12000 documents in a single folder, no metadata, no versioning, nothing fancy. Performance whitepaper aside, 12000 documents shouldn't be a problem, I don't want to hear about it—this is a bug. And meanwhile, from the other end I'm attempting to fend off a total migration of terabytes of network file shares to SharePoint, because "this should work, right?"

I don't want to progress with SharePoint, if my energy is going to be spent tackling the wrong challenges.

Add in the cost of learning

Something that I'm maybe not emphasizing well enough is the point that in theory, all things are possible in SharePoint. And if the cost of becoming a SharePoint expert was zero, SharePoint development would be no question, worlds better than coding by hand. No contest, no question, better.

But learning isn't free.

SharePoint gold rush

Dux Sy refers to a SharePoint gold rush. It's real. It's real in part because the time cost of learning SharePoint is so high. Chris O'Brien (I can't find the entry on his blog) Shane Young states that you should be wary of hiring developers if less than 5 of their last 10 projects are SharePoint projects. It's expensive to find people with that kind of SharePoint experience. Thus, high prices, thus, gold rush.

This equation works for all Enterprise solutions; when I talk about how SharePoint is aggravating my inner engineering brain, trust me, I've worked with other Enterprise systems, and they're worse. SharePoint's a great "Enterprise" product—believe it or not, Microsoft is quite open compared to other vendors, who require all independent consultants be licensed by the vendor, or who provide no public documentation at all—all these extra hurdles artificially inflating consulting prices. Microsoft thankfully doesn't do these things (at least that I notice). So, it could be worse.

No more SharePoint on my blog

I've posted exclusively SharePoint content for a while now, and as I do so, I notice my attitude continues to trend negative. It's my doom-and-gloom engineering brain, it can't just let it go. So, instead of ranting further, I'll just, hmm, let it go. No more SharePoint 2007-related topics on my blog, even if a potential topic is brilliant (WHICH IT INEVITABLY WOULD BE!). No more SharePoint content unless I just can't hold it in.

If you've subscribed to my blog exclusively for the SharePoint bits, sorry; now is the time to unsubscribe.

I'm also officially done learning SharePoint 2007 development, permanently—no more off-hours learning, no more digging through my humongous SharePoint blogroll. When SharePoint 14 is announced, I'll evaluate it to see if they've fixed anything—to see if the situation's improved any. If so, great! Maybe SharePoint 14 will realize the potential of what some are calling "the first real application development platform from Microsoft." There are some awesome business problems that SharePoint already attempts to solve—that with just a little extra help, SharePoint could be a powerful solution. It could be dominant, even.

But if SharePoint 14 is more of the same; if they announce SharePoint Designer Designer 2009, and we're all given the "look what you can do without coding!" demos, and the BDC is still 14 pages of XML without proper tooling, and pieces of the framework are "secretly known" to not work at launch (like the PRIME API), and SharePoint still produces noncompliant HTML out of the box, and I'm still handwriting Feature manifests, and I'm still handwriting CAML, and I'm staring at a doubly-XML-encoded internal field name passed as an argument to a custom XSL function emitting HTML and all this is wrapped in a XML-encoded WebPart tag, and a new SPDataGridView is released and it adds 50 more undocumented methods and properties, and we need to rewrite all our SharePoint 2007 customizations just to make them work with v14…well, we'll see. If MSDN rolls out a SharePoint 14 resource center and it has 3 articles and 50000 useless Sandcastle-generated stub pages—gold rush or no—it may be time to bail.

So what are you going to focus on, Peter?

I have three words for you.

FUN.
DUH.
MENTALS!

Project management (not getting a PMI cert; instead, the meat of project management). Proper object-oriented design. Programming languages (plural). HTML/CSS/JavaScript (they're not going anywhere).

Personal projects!

Programming for fun—something that may even inspire me!

And hey, maybe SharePoint 14 will be a pleasure to work with; maybe I'll jump on that bandwagon…I'm the guy who owns sharepoint14.com, after all.