Caffeinated Geek.ca

Waste….

gibble — Sat, 10 Dec 2011 00:09:08 +0000

I talked about it in my last post, and it still permeates. The Kanban board I have pushed for at work is still not up. The CIO has even approved it. And you’d think a $100 purchase of a whiteboard and some post-its would be a trivially simple task. However, after a couple weeks, still nothing. We ran a trial on a whiteboard in another workers office while they were away on vacation for a week. And you could quickly see where our bottlenecks were. But the team seeing them, and the people who can actually institute change are two different things. Maybe in another couple weeks we’ll finally get it up. Maybe not.

Another fun part of the past few weeks was taking a week to decide on the font for the title of a corporate report. You’d think we would have a standard format. And you’d be right. We do. I followed it initially. Was asked to change it twice since then. Finally had enough and got a PM, BA, and two other employees, and a member of Senior Management. And yet it took that long for someone to explain what the font should be. Turns out, it was exactly like the corporate standard…but no italics. That was it. A week and probably 10-15 man hours between several people to say “No italics”.

Next topic is going to be about UI…and why you don’t let the users design them. Not that they aren’t critical to the process. After all, they’re the owner of the product!

Eliminating Waste

gibble — Thu, 24 Nov 2011 00:26:36 +0000

Like too many projects we start in life, this site became forgotten. Barely started, and with good intentions, I just didn’t find the time to follow through. I have many ideas in my head for projects I want to start…but sadly, that’s how I think of them. As things I want to start. Not as things I want to finish.

A recent talk by Joel Semeniuk at the Prairie Developer Conference in Winnipeg, MB got me thinking about waste. The talk at the time was about Kanban, and how anything started, and not finished, is waste. Unmerged branches, projects not yet deployed, or anything else that’s not “Done. Done.”

I remembered this site. This thing I started a couple years ago. And I realized…it was waste. It’s unfinished. I read and enjoy great blogs by other developers. They help provide insight, solutions, and sometimes a spark. A spark that puts you in the zone. Even if only for a little while, but it motivates you. It makes you strive for perfection.

Which ties into reviews. I, like most developers, find reviews pointless. Along with timesheets, but that’s another rant. This review, in my mind was a giant waste of time. It wasn’t long, it was basically a quick survey where you evaluate yourself. Then later you sit with your boss and go over it. However, what I consider “meets job expectations” and what my boss may consider “meets job expectations” are likely different. And it has been pointed out before that they are, that I was “above average” or “exceeding”. But how is this helping either of us? Should I do less? No…that’s probably not the point. Does it actually help me become better? No. I strive to improve on my own. And if you’re in IT and not striving to learn and improve, you’re in the wrong industry. So if our views were so different, and he already knew how he would rank my work, why were we going through exercise? I really don’t know. Seems like a lot of waste when you calculate the number of employees doing this, and their bosses time, and what could be getting done during it. It seems we only do these because HR says we should. Because they help some employees…so it must help everyone right? Or at the very least, not hurt. All wasted time and effort for little to no business value. At best it breaks even, more thank likely it’s a cost to perform and store these that never provides any overall benefit to the company.

Ultimately, I’m hoping in the next couple days the team and I get a Kanban board up on the wall, with WIP limits (that’s going to be a hard sell) and that we use and enforce the proper flow of tasks through it. I guess we’ll see how it goes.

To sum up, the point of this 500 word rant is that we need to eliminate waste, even if it’s one small step at a time. And I’m going to try and turn this waste of a site into something that can provide me value. Somewhere that I can reflect, and hopefully as a tool to better myself.

BB Tips

gibble — Tue, 27 Oct 2009 22:46:00 +0000

ALT-CAP-H will bring up your “Help Me!” page with pin, platform, version, mac, signal, etc
Typing “MYPIN” in a message will enter your pin for you.
ALT-”NMLL” will toggle between signal strength bars and numbers
ALT-CAP-DEL will reboot your blackberry

Gone Mobile

gibble — Fri, 12 Jun 2009 05:14:27 +0000

The site has gone mobile. It’s looking good on the BB, in fact, that’s where this post is being written. How’s it work from your mobile device?

Wrap existing .NET site in a Portal/Theme

gibble — Thu, 11 Jun 2009 23:17:29 +0000

Recently I ran into the problem where we were attempting to have all of our websites run inside of a portal. Where you could easily jump from one app to another, without having to go back and forth, or always back to a home page with links.

What we wanted was a header menu added to ALL of our websites that handled some authentication, and showed what websites you had access to, and allowed you to easily navigate between them, as though they were one large application.

Now, rewriting all these applications would be a nightmare. There was no time. Even if the goal was simply to modify the master pages of each app…some of which were very dated, and didn’t have the luxury of master pages.

So with a little research on the Page Adapter Class and some creative use of the .browsers file, I came up with an unobtrusive method to easily add our new portal to all existing applications, without any recompiling.

First, lets create our Portal Page Adapter project. Add a new class file named PortalPageAdapter.cs

namespace PortalPageAdapter
{
	using System;
	using System.Collections.Generic;
	using System.Linq;
	using System.Text;
	using System.Web.UI;
	using System.Web.UI.HtmlControls;
	using System.Xml;
	using System.Web.UI.WebControls;
	using System.Web;
	using System.Security.Principal;
	using System.Xml.Linq;
 
	public class PortalPageAdapter : System.Web.UI.Adapters.PageAdapter
	{
		protected override void OnInit(EventArgs e)
		{
			// Do whatever security checks you need to do...
			string user = HttpContext.Current.Request.LogonUserIdentity.Name;
 
			List<string> roles = new List<string>();
 
			//populate your roles
 
			// Setup Context
			HttpContext.Current.User = new GenericPrincipal(HttpContext.Current.User.Identity, roles.Distinct().ToArray());
 
			base.OnInit(e);
		}
 
		protected override void CreateChildControls()
		{
			base.CreateChildControls();
 
			// Inject portal onto page
			HtmlForm html = this.Page.Controls.OfType<HtmlForm>().First();
			html.Controls.AddAt(0, new LiteralControl("
PORTAL GOES HERE"));
 
			// Hide WebControls based on their Role Attribute
			// 	This allows the portal, to handle the rendering of any websites 
			//	you have, it can show/hide items at the control level
 
			List<WebControl> controls = html.Controls.OfType<WebControl>()
				.Where(c => ! string.IsNullOrEmpty(c.Attributes["role"]))
				.ToList();
 
			foreach (WebControl control in controls)
			{
				control.Visible = false;
 
				foreach (string role in control.Attributes["role"].Split(' '))
				{
					if (HttpContext.Current.User.IsInRole(role))
					{
						control.Visible = true;
						break;
					}
				}				
			}
		}
	}
}

Now that we have our portal written, with a trick up it’s sleeve to hide/show content based upon the users roles, let’s look at how we add it to an existing website.

In the App_Browsers folder, add a file called default.browsers

>
	 refID="Default">
		>
			 controlType="System.Web.UI.Page" adapterType="PortalPageAdapter.PortalPageAdapter" />
		>
	>
>

And add the compiled dll of our PortalPageAdapter class to the bin folder.

Navigate to your website and the first thing you see should be the line “PORTAL GOES HERE”, which you can modify in the Portal Page Adapter class to create whatever content you would like. How you generate the html is up to you. Be it includes, databases, xml/xsl, whatever. Get creative.

Now, the last thing, is, the ‘trick’ that has been programmed in, on any aspx page, you can add the “role” attribute to any WebControl (asp:whatever) and it will be rendered (or not) based upon the roles loaded by the PortalPageAdapter class.

<%@ Page Language="C#" AutoEventWireup="true" CodeBehind="Default.aspx.cs" Inherits="PageAdapter._Default" %>

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Vestibulum imperdiet justo id sem. Aenean convallis mi sed ipsum. Donec semper dapibus erat. Suspendisse consequat libero quis felis feugiat aliquam. Sed nec elit. Fusce sapien tellus, vestibulum id, pharetra in, scelerisque sit amet, lectus. Vivamus eu turpis at nunc elementum ultricies. Morbi feugiat fringilla est. Fusce urna diam, accumsan vitae, ultrices non, varius vitae, quam. Vivamus vitae enim vel nulla mattis aliquam. Suspendisse lacinia arcu non urna. Praesent convallis ante ut sapien. Nunc sit amet mauris ornare mauris accumsan ultricies. Praesent venenatis tellus id quam. Fusce vel enim a est malesuada venenatis.

Sed eget mauris vitae libero imperdiet malesuada. Suspendisse feugiat semper erat. Duis sit amet odio ultricies dui ultrices volutpat. Sed id risus vitae odio bibendum varius. Nam rutrum consectetur felis. Donec vitae velit. Phasellus facilisis ornare mi. Etiam quis dui id leo bibendum vehicula. Proin egestas, neque quis pulvinar malesuada, odio tellus porta orci, nec dictum est nunc id diam. Integer eu dui. Vestibulum ante ipsum primis in faucibus orci luctus et ultrices posuere cubilia Curae; Ut nec ante ut pede ornare ornare. Mauris eget mauris et urna tincidunt tincidunt. Suspendisse sit amet lacus in dui fermentum aliquam. Lorem ipsum dolor sit amet, consectetur adipiscing elit. Suspendisse odio. Phasellus orci. Aliquam erat volutpat. Nam eget ligula id arcu fermentum lobortis. Sed vel leo.

Suspendisse pulvinar. Etiam ipsum. Proin vulputate nibh et purus. Nunc hendrerit. Morbi consequat nibh id tortor. In mollis rhoncus velit. Pellentesque magna ipsum, cursus sit amet, viverra eu, porttitor et, dolor. Cras quis sem. Ut interdum nisi quis magna. Donec mauris erat, consequat id, gravida eu, scelerisque ac, nisi. Nulla ante purus, dignissim nec, mollis quis, euismod quis, arcu. In vel arcu. Duis enim erat, accumsan vel, blandit vel, ultricies id, pede. Nulla luctus, nisi at faucibus tincidunt, orci sapien ornare erat, vitae iaculis justo risus ut magna.

Etiam imperdiet lacus at dui fringilla dictum. Etiam condimentum, diam vitae fringilla faucibus, ante ante facilisis nulla, at porttitor diam ipsum vel odio. Aliquam sollicitudin neque et diam. Aenean feugiat, justo ut imperdiet cursus, sem risus ultrices ligula, eget rutrum orci leo tempor enim. Nunc eget velit. Pellentesque dictum, odio a tempor aliquet, ipsum justo lacinia mauris, sed posuere metus enim vel ipsum. Donec dolor. Aliquam semper, eros vitae vehicula lacinia, dolor massa suscipit libero, a molestie enim justo tincidunt sem. Praesent eget mi ut purus interdum vehicula. Nulla a leo. Morbi lacinia, ligula id pharetra vulputate, sapien arcu dignissim tortor, ac condimentum velit libero at purus.

Donec adipiscing lacus ac sapien. Nulla ac quam. Sed enim. Curabitur magna. Cum sociis natoque penatibus et magnis dis parturient montes, nascetur ridiculus mus. Cras viverra, libero quis molestie cursus, justo lacus semper nisl, in fringilla enim pede ac elit. Pellentesque facilisis. Etiam fringilla adipiscing mauris. Vivamus bibendum nibh nec massa. Maecenas risus. Cras tempus accumsan pede. Cras dictum hendrerit dolor. Curabitur et tortor ullamcorper elit mattis lacinia. Cras tempus euismod velit. Donec feugiat nunc quis dui. Vivamus nisi urna, bibendum et, imperdiet sit amet, facilisis in, libero. Maecenas lobortis velit at ante. Aenean in lacus et massa fringilla molestie. Praesent mollis nibh ut nunc.

This would also be a simple way to add an embeded IM client to a series of websites.

Happy coding.

The C# ?? null coalescing operator

gibble — Thu, 11 Jun 2009 04:09:12 +0000

A simple tip to save typing and increase your codes readability is the “double question mark operator”, more accurately called the “null coalescing operator”.

Instead of using this to set defaults in a function

function void test(int aVar) {
	int myVar = aVar;
	if (aVar == null) {
		myVar = 42;
	}
}

function void test(int aVar) {
	int myVar = aVar == null ? 42 : aVar;
}

You can simply use

function void test(int aVar) {
	int myVar = aVar ?? 42;
}

Granted, you’d likely find much better ways to use this than simply defaults in a function call, but hey, this is just an example, let your imagination do the work.

MSDN reference: http://msdn.microsoft.com/en-us/library/ms173224.aspx

XML Validation with XSDs

gibble — Thu, 11 Jun 2009 00:31:42 +0000

I’ve been doing a lot of work with excel uploads lately to allow clients to easily upload data to their systems. Not the best approach mind you, but they know how to use excel…so what can you do?

What I have worked out is a multi step process to import the excel file to the database and validate the contents to a reasonable degree. I have these all written into some custom classes and use IOC and DI to handle workflow so that writing subsequent uploads is a trivial task, and the majority of code is simply the validation and transformation files (XSDs, XSLs, and Stored Procs)

First, you have to retrieve the contents of the excel sheet(s)

protected XElement GetWorksheetXML(int sheetNumber, string xsltPath)
{
  // Create working file
  string workingFilePath = Path.GetTempFileName();
  File.WriteAllBytes(workingFilePath, this.filebytes);
 
  // Read the worksheet
  DataSet myDataSet = new DataSet();
  string strConn = "Provider=Microsoft.Jet.OLEDB.4.0;Data Source=" + workingFilePath + ";Extended Properties='Excel 8.0;HDR=No;IMEX=1;'";
 
  // Fill the Dataset
  OleDbDataAdapter myCommand = new OleDbDataAdapter("SELECT * FROM [" + GetExcelSheetNames(strConn)[sheetNumber] + "]", strConn);
  myCommand.Fill(myDataSet);
 
  // Clean up
  File.Delete(workingFilePath);
 
  // Return transformed XML
  return XElement.Parse(Utility.Transform(myDataSet.GetXml(), xsltPath, null));
}

Secondly, convert it to a better XML format with an XSLT. You should also merge any sheets here so you have one XML file from the XSL. I’m not going to bother showing this, as it’s hardly the point of the article, though, if you are unsure how to do this, feel free to ask

Third, and the whole point of this article, verify data types with XSD.

protected XElement ValidateXSD(XElement item, string xsdPath)
{
  XElement errors = new XElement("Errors");
 
  if (File.Exists(xsdPath))
  {
    // Reference: http://msdn.microsoft.com/en-us/library/bb358456.aspx
    XDocument xsd = XDocument.Load(xsdPath);
    XDocument xml = XDocument.Parse(item.ToString());
    XmlSchemaSet schemas = new XmlSchemaSet();
    schemas.Add("", XmlReader.Create(new StringReader(xsd.ToString())));
    xml.Document.Validate(schemas,
      // Validation Event/Error Handling
      (sender, e) =&gt;
      {
        errors.Add(new XElement("Error", e.Message));
      }
    );
  }
 
  // If there were errors return them, otherwise return null
  return errors.Elements().Count() &gt; 0 ? errors : null;
}

Fourth, at this point, if there are no errors, you should pass the XML file as input to a stored procedure to further validate the data (if there are IDs and Codes that need to be verified with other systems), and assuming all has gone well, pass the XML file to a stored procedure which writes the data into your database…or whatever you need to do. By this point your XML should be clean, and valid.

It’s scary because it’s true…

gibble — Thu, 04 Jun 2009 02:57:37 +0000

The Vendor-Client Relationship

http://www.youtube.com/watch?v=R2a8TRSgzZY

Painlessly Audit Your SQL2005 Schema & DDL Information

gibble — Sat, 04 Apr 2009 06:51:19 +0000

I recently stumbled across the following page

Painlessly Audit Your SQL2005 Schema & DDL Information.

We all know about SQL DML (Data Manipulation Language) triggers, these are the kind we use occassionally when data is manipulated.

What you may not know about are DDL (Data Definition Language) triggers.

In short, you can write a trigger to fire whenever tables, stored procedures, views, etc change.

So with the following table and trigger (code stolen from link above) you can audit your stored proc and table changes

CREATE TABLE [dbo].[DBChangeLog](
    [LogId] [INT] IDENTITY(1,1) NOT NULL,
    [DatabaseName] [VARCHAR](256) NOT NULL,
    [EventType] [VARCHAR](50) NOT NULL,
    [ObjectName] [VARCHAR](256) NOT NULL,
    [ObjectType] [VARCHAR](25) NOT NULL,
    [SqlCommandXml] XML NOT NULL,
    [EventDate] [datetime] NOT NULL CONSTRAINT [DF_EventsLog_EventDate]  DEFAULT (getdate()),
    [LoginName] [VARCHAR](256) NOT NULL
) 
 
CREATE TRIGGER [LOG_DB_OBJECT_CHANGE] ON DATABASE
FOR create_procedure, alter_procedure, drop_procedure
    , create_table, alter_table, drop_table
    , create_function, alter_function, drop_function
AS
 
    SET NOCOUNT ON
 
    DECLARE @DATA XML
    SET @DATA = EVENTDATA()
 
    INSERT INTO DBChangeLog(databasename, eventtype, objectname,
                            objecttype, sqlcommandxml, loginname)
    SELECT @DATA.VALUE('(/EVENT_INSTANCE/DatabaseName)[1]', 'varchar(256)')
        , @DATA.VALUE('(/EVENT_INSTANCE/EventType)[1]', 'varchar(50)')
        , @DATA.VALUE('(/EVENT_INSTANCE/ObjectName)[1]', 'varchar(256)')
        , @DATA.VALUE('(/EVENT_INSTANCE/ObjectType)[1]', 'varchar(25)')
        , @DATA.query('(/EVENT_INSTANCE/TSQLCommand)[1]')
        , @DATA.VALUE('(/EVENT_INSTANCE/LoginName)[1]', 'varchar(256)')
 
GO
SET ANSI_NULLS OFF
GO
SET QUOTED_IDENTIFIER OFF
GO
ENABLE TRIGGER [LOG_DB_OBJECT_CHANGE] ON DATABASE

And that’s all there is to it.

You could now easily write a webpage to search and filter this table, and perform diffs on different versions. And never again have a stored procedure get changed on you, clobbering a fix you did the day before…without having a backup of your fix!

And you can log who made the change, so you can exact some revenge.

-c

Regexes and Named Capture Groups

gibble — Sat, 04 Apr 2009 06:40:40 +0000

Not everyone knows about named capture groups, not everyone knows about regexes for that matter, but for those that do, capture groups are great, they allow us to grab portions of our regular expression match, and use them in our application. The problem is, as your regex gets more and more complicated, if you still rely on the index of the capture group, to get the data you want, it can change as we make modifications, and be hard to find, and introduce undesired behaviours when you set a variable to capture group 7, when it’s now capture group 8…but you didn’t notice. Yeah, it should be caught in testing, but it’s nearly 2009, we shouldn’t be referencing data by numbers when we can reference it by name.

So let’s say we have a flat file that we are parsing, it’s an old data file from a legacy system that we are trying to read in the format

id[tab]name[tab]position

And we’d write a simple regex to match and capture the 3 parts /^(\w+)\t(\w+)\t(\w+)$/

var id = match.Group[1];

var name = match.Group[2];

var position = match.Group[3];

But, let’s say that the file changes slightly and we insert extension between name and position, our regex would change to /^(\w+)\t(\w+)\t(\w+)\t(\w+)$/ and we’d have to change position to use group 4.

A small inconvenience, but unnecessary.

if our matches are named /^(?\w+)\t(?\w+)\t(?\w+)\t(?\w+)$/

We can use match.Group["id"] to get the name, match.Group["name"], etc

In more complex scenarios this saves a LOT of time and headaches. Not only that, it comments your regex making it easier to maintain and read in the future!

-c