Once upon a time I was something of a UNIX expert, but this was in the days before internet security was a threat. (I really never had to worry about malicious hackers and I am essentially naive about protecting systems from such.)
I built scripts in all the various shells that did a lot of my work for me. In a pinch I could write C but I preferred scripts.
My proficiency (and a Master's in Management) got me promoted to manager where I managed a small group of test engineers. As a manager my goal was to get the most work done I could with my
group. However, engineers don't like to take orders, they like to solve problems. So, we would collectively decide on ways to solve the problems that kept us from testing more efficiently. One of the problems we encountered was a way to log our testing results which, due to a change in the corporate structure, needed to be logged via a web site.
I figured out I could do something a lot easier if I had a way to do automated scripting of my navigation of web-pages. Someone recommended we try Lynx and, because none of my engineers were particularly interested in doing it, I did it. (One of the pains of being a hands-on manager/group member is that you keep your engineers working on stuff that they like, that inspires them, and when nobody is inspired by a particular task you take it on yourself.)
I don't have a lot of patience to begin with (probably suffer from ADD) but I _DID_ want to circumvent this logjam in our automated
testing/reporting process. Thus, I read up on Lynx only as much as I needed to in order to solve the problem. My recollection (of about 6 years ago) is that I read enough to learn you could "send" Lynx a series of keystrokes and capture the results (coming from the web page) in a log. I would then parse the log (via script), and, based upon the elements found in the parse send another set of keystrokes.
The upshot of this is I was able to create a "robotic" web navigator that was somewhat intelligent. When we plugged this into our automated testing process it solved our problem. We actually had a testing system which we could schedule tests to occur one after the other and the whole process ran unattended. We even had serial console monitoring so that if the system crashed we could detect it and restart the system. All of this we created from the ground up (though we borrowed workable pieces of script from company archives
when we could).
Back to Lynx - This was pretty easy for me once the sys admin had configured Lynx knowing how I wanted to use it. My recollection was that some features needed to be customized, but the sysadmin was savvy in Lynx and knew what to do. He tuned it and I built the scripts and incorporated them into my testing system.
I was eventually downsized (despite the fact that my group's automated tool chest enabled me to produce 42% of the testing results with a mere 5% of the testing staff) and haven't touched this stuff in years. In the interim I have not worked in technology but I still "play" with it a little.
Recently I was "playing" and loaded FreeBSD on an old desktop. When I was playing I found out that Lynx was there, and tried to play with it. My recollection is that when I tried www.yahoo.com, that worked kinda okay, but logging into the customized my.yahoo.com did not work. Because I was "playing" and
didn't have time to read up on sys admin stuff (and figure it out) I just dropped it.
Now, however, I'm rekindling my old interest in automated system testing and web browsing. Frankly, I've been reading up on web testing strategies and I see all this stuff about WinRunner, LoadRunner, TestDirector, etc., but never do I see Lynx mentioned. This baffles me. If Lynx still works the way it did, it is an EXTREMELY powerful web testing tool. Further, since you can disable images, you can test at much higher speeds over limited band widths.
So, why the heck isn't Lynx touted as the savior of American testing technology? Why is everyone sending testing jobs to India and China when a couple of good engineers running Lynx with scripts can do the work of hundreds of people without the language barrier?
If Lynx has been maintained and can be configured to do web browsing in a robotic sort of way, I believe, using it as I once did,
that it would be a SUPER web testing tool.
What I'm not sure about is whether the evolution in web-pages, java/_javascript_, etc. has resulted in changes that make Lynx unwieldy.
Frankly, if there's a devoted developer/user group and the product still works I'd like to see it get promoted. Actually, heck, as I think about it, I'm enthusiastic enough that I might just be willing to commit to doing some of the promoting... (Unfortunately, as per the above, my code expertise is slim...)
Anyway, is there some sort of core die-hard group that I can get plugged into that will tolerate my lack of coding expertise? Are there any meetings I could attend?