You know, I used to use "net send" to annoy other members of the tech architecture team when we were bored. A batch file makes it nearly impossible to get anything done, which in the right environment will start a nasty rubber band war. But I never extended it to spam.
The question is, how long before some ad service decides to take advantage of this?
I recently migrated to Apple's Mail program but found it's built-in search feature to be about as useful as a 3-day hangover. Basically, I kept trying to use it as I use Gooogle, placing phrases in quotes, separating inclusive AND key words with a comma or plus. But Mail, in it's eagerness to please ignored these logic queues.
However, I have found what I think is a good solution to this problem (aside from moving back to Entourage) in the form of a small Java jar called ZoE. This free tool is at once a mail client, mail server, mail filter, and mail search engine. Just import your existing mailbox, sign up for a Google's Web APIs, and let Zoe Google-ize your mail repository. It is, in a word, slick.
Hi folks. We have just posted the first installment of our October 10th issue. With that, we have the complete results of our Reader Survey. Based on the 2,647 responses we discovered what it's really like to work in IT today --from what it takes to elicit funding for new technologies to keeping users happy to using office politics to your advantage. We also reveal the most highly anticipated trends for 2003.
To help you get started with our newly launched Performance Portal, we've posted here a terrific primer, written by Randall Kennedy. After you've gone through his introduction, we invite you to read his tactical how-to on setting up and testing database servers with the Performance Portal. You can find it on the home page here. Also, if you have any questions for Randall, please don't hesitate to ask.
Enterprise Testing on the Cheap!
Getting Results with the (FREE) Network Computing Performance Portal Service
By Randall C. Kennedy
Where have all the "freebies" gone? These days you can hardly click a hyperlink without running into another pay-as-you-go Web service. Whether it's Apple Computer and the now infamous iTools bait/switch fiasco or Microsoft Corporation and its Hotmail storage extortion schemes ("pay-up or we nuke your messages"), it seems like everyone's trying to erect their own private "tool booth along the information superhighway."
So it should come as a refreshing surprise to learn that Network Computing is introducing a FREE (as in NO charge = gratis = complimentary) enterprise testing solution. The Network Computing Performance Portal service is a versatile collection of client/server load simulation objects and test scripts that have been woven together to create the industry's first fully-integrated online testing environment. Covering a range of enterprise application types (Client/Server Database, Workflow, Multimedia and Productivity), the Performance Portals make it easy for IT organizations to evaluate new hardware/software purchases and analyze existing systems to isolate bottlenecks, identify problem areas and chart performance over time.
Think of the Performance Portals as a set of scalable testing "end points." Through a combination of client-side scripting and robust multi-instancing support, each Portal provides a flexible workload generation capability that is both client-driven and entirely self-contained. Need to stress test a particular network or server resource? Simply login to the Portal site from a client PC, configure the appropriate workload package and click "Start." The Performance Portal service's "zero server footprint" design means no messy test code running on your critical systems. You can safely use the Portal workload objects to analyze live production environments without risking downtime due to a failed test controller or similar plumbing issue.
Did we mention it's FREE? You can use the Performance Portal service as much as you like. Run it on as many systems as you need to support your project goals. Go nuts! Fire up that room full of idle PCs, log them all into your Portal account and turn your humble enterprise lab into a client/server testing powerhouse. We'll even host your configuration and results data for you at our web site, allowing you to access your critical information from any web-enabled PC, anywhere in the world. Just surf up to the Portal site, login using your unique credentials, and you're right back where you left off. No more forgotten program CDs or missing configuration files - it's the ultimate in portable, personalized testing.
What's it Good For?
To quote the Fabulous Thunderbirds: "It's in the way that you use it." The Performance Portals are generic by design: Though the test components are Microsoft Windows-based (a combination of ActiveX Controls and Servers), they are built around flexible APIs that allow them to interface with a variety of back-end resources.
For example, the Database Portal workload simulator uses the Microsoft ActiveX Data Objects (ADO) libraries to connect to, and interact with, any OLE-DB or ODBC-accessible data device. Since virtually every database solution on the market supports one or both of these protocols, you can be fairly certain that some combination of connection strings will work in your environment. In fact, the entire Database simulator is engineered to be platform agnostic, from the way it sequences transactions to the format of the SQL language scripts it executes.
This inherent flexibility is what makes the Performance Portal components so powerful. In this past year alone, CSA Research - the host of the Portal site and the author of the Benchmark Studio Object Framework on which the site's based - has used the Database Portal to execute contract testing projects for a variety of industry heavy-hitters, including Intel Corporation (Gigabit Ethernet testing, Hyperthreading scalability project) and Rambus, Inc. (server memory bandwidth analysis). Thanks to the sensitivity and all around robustness of the Portal workload components, CSA was able to develop a series of innovative testing packages that helped to prove out some very elusive concepts. You can learn more about these projects by reading the corresponding white papers and project report articles hosted at the Performance Portal site (visit the Library page for direct links).
Of course, these same principles - powerful, client-driven testing of network resources via intelligent end points - can be applied to virtually any enterprise testing scenario. For example, the methodology from the aforementioned Gigabit project can be used to test network performance across any type of link. Here another of the Portal design features - the use of high-level APIs and subsystems - provides an edge. Because the workloads are abstracted from the underlying topology, they can run unmodified across the broadest possible spectrum of connection types and transport technologies. Sometimes it pays to follow the K.I.S.S. principle (Keep it Simple, Stupid)!
Where to Go from Here
The first step is to sign-up for the Performance Portal service. The registration process is straightforward: Just drop by http://perform.networkcomputing.com and follow the link to create your new Portal account. Then login and get to work - it's that simple! Everything you need to be productive, from "How To" guides to sample test projects (including the aforementioned Gigabit report), can be found in the Portal Library or Online Help system.
They say the best things in life are free. In the case of the Performance Portal service, we're taking functionality that would normally cost thousands of dollars to implement (via commercially available solutions) and delivering it as a FREE resource for the greater IT community. Enterprise-caliber performance testing is an integral part of what we do every day at Network Computing. It's time we shared some of that "secret sauce" and helped our readership take their own testing projects to the next level.
- Randall C. Kennedy, Director of Research, Competitive Systems Analysis, Inc.
I don't know about you, but one of the biggest banes of a web wonk's life is the meta tag, that strange, metaphysical entity we all seem to rely upon but never quite trust in first identifying and later finding our precious Web content.
In his article, Danny calls for the dissolution of meta tags, mostly because only one search engine (Inktomi) actually uses the incredulous things in building its indices. While I agree with Danny that meta tags really have a limited use when it comes to positioning your content within a winning search engine, I contend that meta tags must continue, not for external but for internal search engines.
We presently use a host of meta tags in our content, each designed to provide a crucial bit of data regarding our content, such as the date published, the technologies discussed, the author's name, and of course, the story's title. We do this not so much for external engines but for our internal Verity engine, which is tuned to identify and then prioritize specific tags.
For example, we use a tag called "article type" which helps Verity build sub-collections of data for quicker searching and listing. Here's an example:
<meta name="article type" content="Feature"/>
We also use a tag called "FirstPage" which tells verity that a given page is the first page of a multi-page story. We do this simply because we want to exclude a series of search results that appear to be duplicates. Clearly, without custom meta tags like these, our internal search engine would operate at serious disadvantage. However, that's not what I worry about with meta tags.
For the past few years, I've grown increasingly ill at ease with these critters, knowing that such tags fail for one simple reason: people.
How can you rely upon the validity and consistency of tags that are so often created by individuals, who, through no fault of their own, cannot count on or be sure of their ability to enter a simple date the same way each time. Have you ever seen the date encoding? Here's an example right from the WWW Consortium's HTML 4.0 specification:
What we need, and what we've only recently been able to taste here at Network Computing, is a content delivery system that generates meta tags programmatically, putting an end to erroneous and inconsistent meta tag entries. Our editors simply select the story type, publication date, author, etc. from a UI that can verify and correct their entries. We hope this mechanism will eventually lead to a very efficient historical record of all NWC content, one which you can search with confidence.
As far as external engines like Google.com are concerned, we love them and we say, "let them eat links."
