Recent Blog Posts
Blogs ยป Archive for May, 2005
Is The Relevancy Challenge Relevant?
Barry Schwartz of SE Roundtable had an idea to put the relevancy of the major search engines to the test by creating a “white labeled” search engine that would display the results from the different search engines and ask users to rate the relevance of each. (Click that link before Barry gets buried for borrowing six trademarks).
This idea has been criticized and praised in various places, but Barry was undaunted and actually implemented a white labeled “RustySearch” engine. Thus, the great Search Engine Relevancy Challenge of 2005 was born.
Danny Sullivan and others have pointed out that there really isn’t much distance between the major search engines, and early results appear to bear that out. Yahoo has taken an early lead with Google right behind, Ask Jeeves and MSN are 3rd and 4th. The difference between the #1 Yahoo (score 3.38 out of 5) and #4 MSN (score 3.07) is not much to speak of, especially since these results are based on 3,200 searches which were probably performed mostly by SEOs with a fairly narrow range of search terms.
So is the Relevancy Challenge really relevant? Maybe not, but it’s still interesting, and I hope that Barry will continue it for …
A Fresh Look at Open Source
Sometimes all it takes is a step back from a technology sector to have a fresh (or refreshed) perspective.
I have spent the last several weeks immersed in a substantial enterprise project that is one hundred percent not open source. It is the exact opposite - built on a comprehensively proprietary platform from top to bottom. I am pleased to disclose it will do just fine (of course - I was working on it - right!?). However - it also gave me a clearer view of where open source can and will succeed in the enterprise - and the obstacles it faces.
Note: I have been working with open source since 1998 - however - with proprietary platforms since 1992 - and quite honestly both pay the bills. ;>)
Now - being under an NDA (a fancy acronym for ‘keep thine mouth shut about the details’) I will not reveal too much. However, it is a fine project that consolidates numerous data sources across the US into one data mart and will ultimately offer a very nice consolidated web application for reporting and other interesting activities. To boot, the company looks to save several million dollars annually in costs …
When do you give fixed bid?
On larger projects, and especially those with fairly un-savvy clients (technologically speaking), it makes sense to go with time and materials bid.
But there are times to give a fixed bid:
- Simple projects where it is easy to estimate time and materials, and there is little risk of conflict over scope or the definition of “done.”
- Projects when you want to get a foot in the door and show what you can do, even if you may only break even (e.g. client could be very valuable long term). Best to do this on small, niche projects.
- When you have packaged your solution in a repeatable way, and so know what you need to spend to provide results.
- When certain audiences demand it. For instance, many construction firms are used to providing fixed bids, and insist that certain IT vendors do same. But here, you have to be darn sure you know what you are getting into before you go after that kind of market.
Do you agree or disagree?
Ho, Hum, another Google patent…
No, not another new one today, sorry. In fact, this patent, for Information retrieval based on historical data, has already been discussed to death. A nice example can be found at SEOmoz.
So, why haven’t I blogged it to death myself? Because it tells us nothing about what Google is doing today, and very little about what they might do in the future. Read the SEOmoz discussion and the patent, and you’ll see that the patent covers just about every possible way that historical data might be collected and used.
If Google decides to reward new sites which acquire a lot of backlinks in a hurry, because great new resources tend to get a lot of links in a hurry, that would be covered under this patent. If Google decides to penalize new sites which acquire a lot of backlinks in a hurry, because that’s also what link spammers do, that would be covered under this patent.
So it goes with just about any historical data they might look at… the problem is discerning the “good guy” cases from the “bad guy” cases. Even though Google’s decision makers may have no idea what they want to do, they file a patent …
Zend Platform (and a free T-shirt)
Zend is giving away free T-shirts to anyone willing to complete a pre-evalutation survey and commence a free evaluation of Zend Platform. This news comes via PHPDeveloper.org.
From the product information:
Through a centralized management console that proactively dispatches information to the applicable IT personnel, Zend Platform delivers comprehensive insight, run-time profiling, and performance monitoring, with an unmatched performance boost to ensure your business runs optimally.
The product can send automated alerts to the relevant employee when errors occur, and it logs information about the errors. A live demo demonstrates the user interface, showing statistics about incidences such as slow queries, peaks in server load, regular PHP errors, and inconsistent output sizes which could be indicative of other problems.
The software costs $995 per year for a single CPU license (a bit like adding $83 per month to your dedicated server costs).
Link research (and blogging) made easier
My buddy Aaron Wall, who writes a little blog called the SEO Book, has released a very nice tool for doing link research called Hub Finder. As far as I know, nobody’s currently hosting this tool online, but you can download the source code and install it yourself. The purpose of this application is to find topically related sites by looking at common backlinks between URLs. To quote Aaron, “none of the major text link analysis tools for sale allow you to check co-citation, or pages which link to multiple related resources.” This one does. :D
Aaron has recently launched another tool called Link Harvester which does link searches beyond the 1000 link limit normally returned by search engines. It is available here, for download or to try online.
Both of these applications are Open Source, GPL, and free. I’ve got my own link research application in development similar to Hub Finder, which will also be released under the GPL terms.
Thanks to Aaron (and the programmers involved) for contributing to the pool of useful source code, and for giving me something to blog about on a slow news day.
Coming software shakeout?
http://www.fortune.com/fortune/fastforward/0,15704,1058043,00.html?cnn=yes
The above link takes you to a good article in Fortune about the state of the large-scale software industry. This article shows tremendous opportunities for smaller players.
Basically, companies are fed up with expensive, time-consuming, one-size-fits-all installations that get no results.
They also aren’t sold on open source solutions.
What they want is software as a service, not a product, something they can customize and that works out of the box. Salesforce.com is given as a great example.
To me, this points to big opportunities for small to mid-sized developers who can come up with modular, configurable, flexible solutions.
The one caveat is that an increasing number of companies are so frustrated by ‘abysmal’ external help that they are writing their own code in-house.
Radical Interface Approaches
Improvement is usually an incremental thing.
Look at your last two cars. Your last two cell phones. Smart people take the current best version of something and make a small adjustment to it — keeping all the other nice things that worked before. Baby and the bathwater.
But not always. Sometimes when the playing field evolves, radical invention get a slight edge.
Web interface design hasn’t changed a lot in the last 5 years. Sure, the coding techniques underlying it have evolved, but the visual structures are pretty similar — header, footer, column layouts, ‘teaser’ content leading to longer articles, etc.
However, the environment that our interfaces are being lauched into has changed. Users are more sophisticated, have faster connections, faster computers and smarter browsers (in relative terms) — all of which presents opportunities to try interfaces that wouldn’t have been possible 5 years ago.
The two following examples have been around for a little while but, I think, are worth spotlighting.
Case 1: Newsmap: Newsmap is a radical approach to understanding the current state of the world through Google News.
Stories on the same subject are grouped and ‘weighted’ on their prevalence — widely-reported stories have larger boxes. Stories …
Moment of truth, continued
In the last blog, about what goes through the prospect’s head (mine, at least) just before choosing an IT Professional, ArtsyTECH posted a great comment:
“On the flip-side, there are many tire-kickers and time-wasters out there that drag negotiations on… and from the perspective of Freelancing, there are many interesting projects and decisive individuals that know what they want when they see it without extended pre-amble… after more than 10 years I’ve mostly found the latter relationships rewarding and the former dissapointing. Procrastination is a serious red-flag for me.”
That’s an excellent summation of setting criteria for ideal prospects, and sticking with your criteria. If you have a solid set of referrals, you can set your definition of a good prospect, and focus your effort on getting hired by them.
To me though, spending a couple extra minutes when a prospect is about to decide — reassuring, showing your interest, making suggestions, confirming whether you have addressed their concerns — is worth the investment of time. You’ve already spent time proposing, and already should have qualified the prospect by this point, so why not do your best to win the engagement you have chosen to pursue?
In diplomatic summary, I believe that ARTSYTech’s comments …
Validate Your Input!
The previous blog post about MySQL and PostgreSQL ignited a discussion about validation of input in a database-driven web application.
In this entry, I’ll attempt to explain what validation is and why it is important.
I remember being taught the importance of input validation way back in high school, in an IT elective subject. We were taught that all user input must be validated prior to acting upon it. At that stage we were ‘programming’ in Microsoft Excel using VBscript, but validation is all the more relevant for web applications, which may be used - or exploited - by a wide range of people.
Data validation is the process of checking user input to see whether it is in the expected format and is within the range of allowed values for that input. It has a number of purposes, some of which are below.
- If a user submits data that is not within the allowed values or it is in the wrong format, it may cause the application to exhibit unexpected behaviour - which may include a blank screen or a screen that doesn’t make sense. Validation allows for this to be prevented, and instead to present a human-readable error …
Sponsored Links
Logo Design, Web page Design and more!
- Custom logo designs created ‘just for you’.
- Pick the design you like best.
- Only pay if you’re satisfied with the result.
The most complete and up-to-date HTML Encyclopedia money can buy.
Learn about this Book
FREE SampleDownload sample chapters of any of our popular books.
