Some Thoughts on Developing Purchase Requirements (for IT Purchases)

You have a business problem that requires some sort of software to solve. There are a number of potential products on the market, but which of them best matches your current and future needs? To successfully solve this problem you first need to define your purchase requirements, and then find the product that best matches those requirements.


Thus your immediate problem is how to develop your purchase requirements. To help describe this requirements development process, I'm going to frame this article around a cloud service purchase. Specifically I'm thinking of software that will be used internally, as opposed to software that will be used by your organization's customers. Examples of this type of internal software are systems for accounting, ERP, a help desk, project management, etc. Note that this requirements development process could just as easily apply to purchasing in-house software, hardware, other services, or any combination of those. A phone system is a good example of a purchase that combines hardware, software, and services.


There are potentially many products on the market that could work, but the question on your mind is this: Which product most closely matches my requirements? How do I find that product? Bear in mind the first order needs for users of these products are often very similar; it is the second order needs that vary significantly between organizations, and they can make or break the purchase.


To answer that question of which product most closely matches your requirements, you need to effectively develop or gather your purchase requirements. But before we get into gathering the requirements themselves, we need a "requirements framework". A list of requirements by itself is of limited use. You also need to know information behind each requirement, basically why it exists, and who it matters to.


Typically you can manage this information on a spreadsheet, and I have found the following columns to be useful:

• Requirement Name - a short, descriptive name for the requirement. Used to refer to the requirement in other documentation.
• Requirement Description - This expands on the requirement name, and is a detailed description of what the requirement is. Write it in such a manner that a potential product vendor would understand exactly what you wanted when reading the description.
• Who wants the requirement - This is a list of those people who want the requirement. It should also include their title or job function.
• Requirement Reason - This is an extremely important field, and explains why the people above want the requirement.
• Requirement Importance - See below for more details on the Requirements Importance column.

Requirements Importance
This is a measure of how important the requirement is to the people who want that requirement. Usually it is best to have words that describe requirement importance, along with an explanation of what those words mean. You can also attach a numerical rating to each importance, e.g. the numbers 0 - 6 in the example below (more on using these numbers in a future post).

Requirement Groups

It is useful to group related requirements in your list, e.g. all security requirements can go in a group called “Security”, reporting requirements can be collected under “Reporting and Analysis” etc. See below for an example of requirement groups from a help desk evaluation.

This sums up a useful framework for managing the purchase requirements you are about to collect before making a significant IT purchase decision. In my next blog article I will discuss a technique for creating a list of requirements that includes finding those requirements you don’t even know you need.

User help in MS Office

A few weeks ago I upgraded to MS Office 2010. In general Microsoft has done a pretty good job with the 2010 release of office. But one thing remains a problem, namely the help engine. In fact, it turns out that if you ask your search question on Google, you get MUCH better results than going through the Office help system.

Online help is a problem with many products; but it is a little unexpected that Google serves up significantly better results that the native tool. It just shows the power of the community. I have long though that all application help systems should be collaborative (and moderated to remove the inevitable trash that collects). You might call it a "crowd sourced" model for  help. But by collecting and harnessing input from real users, the quality of the application help will rapidly improve for the benefit of all users.

Backup software for the home: A review of Mozy

I consider myself to be an IT professional after spending most of my career in IT. A while ago I realized that my backup strategy on my own home PC was not very professional. Much like security, backup is a layered approach, designed to reduce risk. This means you should not just stop at one backup solution. I have a Seagate USB backup drive, but the problem is that if there is something like a house fire, you can lose your backups. Since we keep all our family videos and photos on the PC, the thought of losing all of that in a house fire was just too painful to bear. After some research, I signed up with Mozy in June 08. One of the appeals of Mozy was the unlimited personal backups for just $5 per month because I have over 800 GB of data.

Fast forward about 2 years. A few months ago my XP home computer required rebuilding because it was the easiest way to resolve problems that had developed. It was also an opportunity to increase the size of the hard disk. Also, although I had two backups, the most prudent approach was to buy a new hard drive, and not reformat the old one. Then the unexpected happened – first the Seagate drive demanded that I create new partitions. It has lost all my data! No idea why this happened, but nothing I could do would get my data back. Then when I set up Mozy on the new hard drive, about 75% of my files seemed to have vanished. Although Mozy’s documentation makes moving your Mozy account to a new PC look simple, it really isn’t. In my case, the old hard drive was installed as a 3^rd drive on the PC, and I was able to copy all the files across to the new drive, and no files were lost. And when Mozy was doing the initial backup of the new hard drive, it found most of these files on the Mozy system. Perhaps they were not lost after all, but they certainly weren’t visible in the recovery pane. My suggestion is that if you are moving Mozy from one system to another; do not delete the old system in Mozy. Rather create a new system on Mozy, and only once you have all your files on the new system, delete the old one from Mozy. It will cost you $5 per month extra, but at least you won’t lose any data. While prudence had paid off (i.e not reformatting the old drive), it was a bit sobering to think that both backups had failed.

Web site legal and usability problems

My eldest is in college, and today she asked me to purchase an online textbook from cengagebrain.com. There were several interesting things about this site...

1. At the bottom of the purchasing page there was a link to their terms and conditions - all 8188 words of it! That works out to 260 words per dollar spent. Can you believe it - having to study an 8188 word purchase contract before you buy a digital textbook? Something is not quite right here, especially when the first assignment was due at midnight tomorrow.
2. The last term on the contract stated that they had the right to deny you access to your account at their sole discretion for any reason what so ever. Wow - so I am paying for this access, and if they are in a bad mood tomorrow, they can just turn off my access. I have no recourse at all.
3. Paying for the text book required my credit cards. When entering the credit card number I had to take care to strip out all spaces and dashes - enter numbers only. This is a pet usability peeve of mine - it doesn't take much JavaScript to strip those things out, but no, you the user, must do the work. But if you type one of those 16 digits incorrectly and there are no spaces in the 16 numbers, it can be difficult to find the error. I am never sure if this is just poor UI design, or lazy programmers, but you so often see it on web sites. My rule of thumb is that if the system can do the work, the then user should not have to do it.
4. The credit card expiry date month defaulted to August, but my expiry month for this card is earlier in the year. So when I changed the month I got an error, because the expiry date was now earlier than today (I had not yet set the year field). But the year field is to the right of the month field; naturally you enter the month first. This is an example of poor web site design, but something that can so easily slip past the people developing the web site, and their quality department. These sort of things really take away for the "quality" of the user experience.

It was quite interesting to come up with so many problems with so little interaction on a web site. This is a good example of why work flow, web site & software development really are so hard to get right.

A little on Evernote indexing from images

I have been using Evernote since about March '09, and have found it to work well (4499 notes as of today). I have always known they analyze images and extract the text in the background to index it, but never had occasion to make use of that until yesterday.

While searching for a particular person's name, I saw that Evernote found it on a screen shot that had been pasted into a note. Now that was impressive, because it was an image of the name; there was no actual text there. Evernote recognized the text and indexed it; in fact the text was even highlighted in the image when I searched for the name.

I also use Evernote to store images of all the business cards I get, rather than leaving the cards in a pile in a drawer. Along with the image of the card I usually a few words about the person and why I have their business card. Now I know that Evernote can find all those names in the business cards. Taking the idea a bit further, if you take a photo of a meeting room whiteboard, Evenote will analyze and OCR (Optical Character Recognition) the text, and then index it.

All of this takes Evernote's usefulness up a notch!

Blog move

For some months technical problems have prevented me from posting on my blog; and a lack of time prevented me from fixing them. I have been thinking of using a SaaS (Software as a Service) approach for hosting this blog for quite a while, and these problems pushed me over the edge. After looking at the possibilities available, Google seemed an excellent choice.

I must say I have been very pleasantly surprised by how well Google's Blogger just works.  One example is "After the jump summaries". This is a feature which allows you to create expandable post summaries in your blog posts, so longer posts appear as an intro with a link to "Read More".

There are a few formatting issues that I need to resolve, but that will come later.Once all the articles are moved, I will point the blog's URL over here. The articles will keep their original dates, and if possible I will try to keep the original URLs. Unfortunately the comments probably won't make it across. Once the blog is in its new home it should be significantly easier to maintain, which means I should add articles more regularly again.

Managing information in your personal and professional life

While all of us suffer from an information overload, how we deal with that information is often what separates us on a professional level. This blog entry is a review of my experiences managing this information, leading up to Evernote, a product that works well in practice.

Here at work we are slowly implementing a records management policy, and a probable outcome is that we will be unable to keep email older than one year. As an IT professional this has implications, because my email is full of useful information that I want to keep. After discussing this at length with our records manager, I came to the conclusion that I really need some sort of "personal information manager". In about March 2009 I discovered Evernote, an online service.

Keyword tags and email

In Q2, 2007 I changed companies, which also meant a change to Outlook email. The immediate frustration of Outlook was the inability to put an email message in more that one folder. (You can copy a message to multiple folders, but this duplicates messages). On top of that the Outlook 2003 search is anemic. Upgrading to outlook 2007 helped searching, but the problem of how to file email remained.

Then I discovered an Outlook add-in product called Taglocity which allows you to add keyword tags to messages. After using keyword tags for several months I realized that this approach actually woks better that filing messages in multiple folders. The reason is that you can search on multiple keyword tags. Its a bit like searching for a message that exists in multiple folders. Although that is something I never did with folders, I do it all the time with keyword tags and find it very useful. The only problem with Taglocity is that it is an add on to Outlook, which means that if I use the web interface, I don't have access to my keyword searching.

The interesting thing about using Outlook now for over two and a half years is that it is no better or worse that Notes email; just different. And that although the Exchange back end has been reliable, there are just as many calendar issues with Outlook as there were with Notes. It seems like the calendar is just a very difficult problem to solve.

SharePoint Adventures

My present company is a Microsoft shop and, although we have a very small Notes setup, we decided to use SharePoint as a collaboration tool. So far we have WSS 3.0 up on VMWare, and plan to go to the full MOSS as soon as we have the data center question resolved (more on that in another post). Obviously each product has its weaknesses and strengths, and it has been very interesting comparing the way things are done in SharePoint with the way things are done in Notes. Here are a few little things that I have found in SharePoint that I really like.

New version of DocMaster: 4.2.14

DocMaster is a Single Database Document Management System. Its is used to manage large collections of loosely related documents, for example Policies & Procedures, Application Help, User Manuals, System Documentation etc. Of course the DocMaster User Manual and Application Help are written in DocMaster. Originally developed in 1994, it has been constantly enhanced over the years in response to user suggestions. Here is a link to the original DocMaster posting with more details.

A Thought on SharePoint as File Server 2.0

A while ago Ed Brill had an article in his blog which mentioned SharePoint being described as File Server 2.0 (a phrase coined by Sean Burgess). This got me thinking - originally Vista was going to have a new data storage and management system called WinFS. I had even read somewhere that WinFS was going to be very much like Lotus Notes in the way it handles data. Then in June 2006 Microsoft shelved those WinFS plans. Did they pull the plug because SharePoint was going to fill that need? Basically could they make more money packaging the functionality as SharePoint that they would building WinFS into Windows? Of course there is no way to know one way or the other, but the phrase File Server 2.0 certainly did get me thinking.

Added Document Issues in DocMaster release 4.2.x

What do you do when you are reading though a published document like a policy or procedure, and find a problem or error? DocMaster now has an answer: create an issue. Issues are a special form of comments, complete with a simple (but customizable) workflow. In the screen shot below you can see an example of a new issue, along with the issue categorization and workflow.

Added Sarbanes Oxley (Sox) to DocMaster in release 4.2

One of my responsibilities at work is to ensure that IT is Sox compliant. Unfortunately the usual manual processes based on spreadsheets and word documents simply don't work very well, and are a lot of effort to maintain. And since Sox procedures should be integrated with other IT procedures, it made sense to extend DocMaster to specifically support Sox.

Possibly the most useful feature added is the "Sox Documentation" view, which includes a color coded calendar showing the current status of Sox monitoring. This is a huge time saver, and has organized our Sox work so well that, at the end of May 08, we are completely up to date! See screen shot below with annotations.



Adding monitoring tasks to the calendar is pretty easy: simply check the appropriate month check box on the control document. See screen shot below.



Sox documentation usually consists of control narrative documents, which are summarized by a spreadsheet. The problem is that it is impossible to keep the narrative documents in sync with the spreadsheet. DocMaster solves this problem by using a separate document for each control, and then summarizing these documents in a Control Matrix view. That normalizes the information, and because the information is now only maintained in one place, total solves the synch problem. DocMaster has special views of the Sox control documents, but those control docs can also be put in folders, along with other policies & procedures.

Once Sox controls are in place, you are required to monitor them to ensure they are working. DocMaster allows you to attach monitoring documents to the controls. This means that when the next monitoring event is due, you can use previous monitoring as a template - a tremendous time saver. It also means that when auditors look at your documentation, they can easily find all monitoring - a tremendous money saver! To track actual monitoring there is a view of all monitoring, categorized by month. See screen shot below.



Release 4.2 also added document issues, but more on that tomorrow.

More Excel Woes!

Although the technique "copy as table and paste into spreadsheet" works, it is limited, e.g. if the view is too wide, you can't copy it as a table (but shrinking the column widths does work). Also, graphic icons in a view don't get pasted.

An alternative thought occurred to me: Why not export the view as a Lotus 123 file, and then import it into a IBM Symphony spreadsheet. Surely two IBM products would be compatible with each other? Surely...? Not a chance! The error message is "IBM Symphony do not support this file format." (love the grammar!) Unfortunately Open Office can't import Lotus 123 spreadsheets either, and I haven't yet been able to find any filters on the web. I'll keep trying because there is a need coming up in a few weeks.

Excel import workaround that’s better than the original!

Well, you learn something new every day, and sometimes from the most surprising sources. Although we use Outlook for email, we had created an Asset Management application in Notes (using rich text fields to store scans of original invoices etc). We needed a way to export information from this application to an Excel spreadsheet to further manipulate it. This was where I discovered that Excel 2007 no longer imports Lotus 123 - see blog post from April 23.

A few days later the finance person using the Asset Tracker came back to me and said she had solved the export / import problem - just use "Copy as table" and paste into the spreadsheet! Wow - this finance person had never used Notes before, and come up with a work around that was much better / easier to use than the original process. (Note: Although cell background colors are preserved in Open Office, they are miss aligned. Background colors don't paste into MS Excel.) This just goes to show that you never stop learning, and that potentially anybody can teach you something.

Excel 2007 can’t open Lotus 123 spreadsheets

We hit this problem today trying to export a Notes view as a spreadsheet, and import it into Excel. Older versions of Excel could import a wk4 spreadsheet, but not so in Excel 2007. Maybe IBM / Lotus will finally provide a means to export native Excel files directly from Notes (or maybe this is just wishful thinking...!)

Problems with Microsoft Server 2008?

Microsoft obviously run Windows servers, which they tout as highly reliable. Yet their eOpen web site has had a problem for the past 6 days (and maybe longer - it is just that I have been trying to download some software that we purchased for the past 6 days). I can only presume they are upgrading to Windows Server 2008. But that hardly inspires confidence when considering Windows Server. If Microsoft has this much difficulty getting their software to work, what will I experience?

The strangest thing is that I can reach the eOpen page with Firefox, but not with IE. But that doesn't make any difference, because I can't log in anyway.

Outlook 2007 Calendar Bug

As a long-time Notes user who now uses Outlook at work, I always keep an eye open for problems / differences etc. Here is one problem with Outlook 2007 where if you drag an HTML email to your calendar, you loose the formatting & images. See the before and after images.

The original email before dragging it to the calendar


The calendar entry created from the email

Thoughts on User interfaces - Pages vs Dialogs

I am one of those people who like to keep their financial affairs in order. To help with this task I have used MS Money since about 1993. A while ago problem with Money made me decide to change to Quicken. It was then that I realized the difference between the two classes of interface you find in the software world:  broadly speaking they are dialog boxes versus pages.

Dialog boxes developed out of the character mode world where screen real estate was very limited. These days typical Windows applications have dialog boxes scattered throughout. A dialog box works by popping up over the underlying screen, and displaying relevant, context sensitive information when it is needed. Although dialog boxes work very well in some situations, they tend to be over used. Quicken (2006 version) is a very good example of this. When working with Quicken it is very easy to get into a situation where so many dialog boxes are open that you have no idea of where you are. And at that point the user interface has failed. What also shows up in Quicken is that dialog boxes do not scale. They work well when deployed on a small scale, but quickly break down when they are over used.

Although MS Money uses dialog boxes in many places, they also use pages to control many options.  Because pages are larger, you can get more on them without cramming everything in (just like printed page layout, white space is important to help make sense of everything presented.) But there is something else going on here: it turns out that the paged interface is just easier for people to use. Our brains seem to be wired that way. As evidence I offer the way the internet has caught on. I think a large part of the success of the internet is that people relate to pages much better that they do to dialog boxes.

I spent about three months reviewing Quicken before taking the plunge and changing over. Doing my homework, reading the reviews, deciding that this was the right thing to do, etc. But very soon after converting my files and using it with real data, it became apparent that a huge mistake had been made. Now I have been working with software for over 10 years, so learning a new package should not be that hard. Yet even allowing for the fact that Quicken was a new package, I was taking 2 to 3 times longer to get the same tasks done compared to MS Money. And that was after using Quicken for two months. I can only attribute the difference to the user interface. And a large part of that is the difference between over using dialog boxes and the page interface.

About six months later when Microsoft release their new version of Money that could import Quicken files I swapped back. And, finally, I can balance my loan accounts again (you simply can't do that in Quicken.). And I get my accounts done in less than half the time it too with Quicken.

(Note: In this article I am not reviewing Quicken. I am simply comparing the Quicken user interface to the one in MS Money to illustrate the difference between the two approaches. If you have been using Quicken for years, it obviously works for you.)

Color Blind Spreadsheets

While working on our Sox projects I was emailing spreadsheets around. To help track progress I "traffic lighted" certain cells: Green for no problem, yellow when I was waiting for someone, and red when I had to do something myself. I was using Excel 2007, but those receiving my spreadsheets were using Excel 2003. Every time I saved the file, Excel 2007 would wand me about a minor loss of fidelity but I paid little attention. That was until one user complained that there were no green cells on the spreadsheet – only yellow.

Now I must admit to being color deficient (or partially color blind). So when I chose green for the cells, it was a bit of a yellowish green. And when Excell saved my spreadsheet in 2003 format, it changed these green cells to bright yellow! My guess is that the 2007 version of the program uses 24 bit color (like a photo), whereas the 2003 version uses only 8 bit color (like a gif file). And when the 24 bit color was downgraded to 8 bit, the yellow green was changed to bright yellow – totally altering the meaning!

So beware when using Office 2007 and 2003 products in a mixed environment. Things can bite you if you are not careful.