-
Website
http://mashable.com/ -
Original page
http://mashable.com/2008/07/13/software-development-cycles/ -
Subscribe
All Comments -
Community
-
Top Commenters
-
Robert Basil
142 comments · 8 points
-
Jennifer Van Grove
149 comments · 23 points
-
r0cketman22
317 comments · 52 points
-
rajagiri4
160 comments · 2 points
-
barringtonarch
150 comments · 4 points
-
-
Popular Threads
-
Enter the Zappos Sharing Happiness $3,000 Shopping Spree Giveaway Contest
4 hours ago · 88 comments
-
Your Next Car Radio Might Be Pandora
3 hours ago · 21 comments
-
Google Launches Chrome for Mac
5 hours ago · 26 comments
-
iPhone App Offers Instant Speech-to-Text Transcription
2 hours ago · 16 comments
-
BREAKING: Google Launches Real-Time Search Results
1 day ago · 96 comments
-
Enter the Zappos Sharing Happiness $3,000 Shopping Spree Giveaway Contest
Somewhat joking: I wonder what the "version numbers" of Flickr and Gmail are and have been -- are we in the tens of thousands or have there been an indefinite amount of decimal version numbers?
Yes, it is frustrating to try to understand the development cycle and there is an abuse with terms (i.e. something some people consider "beta" isn't even "alpha" yet and sometimes there's betas that are much better than most people's RC's)... but the lines are blurred and it is only going to get worse, at least as far as web apps go.
Though I guess those lines are blurred when dealing with web apps because most web apps go through changes.
I've actually been meaning to write about why I decided to do a Gridjit Alpha (thanks for the mention!) and your post inspired me to do it today.
Interested readers can view a longer response on the Gridjit blog (http://blog.gridjit.com).
After an internal "dogfood" phase in late March (after just 3 weeks of initial development), we determined that the best way for us to build socialmedian was to gradually expose the early service to more external users while gathering their feedback and rapidly iterating. So, we threw some half-baked software out in front of some folks and hoped for the best. We called it "alpha" because it was invite only and buggy and in many ways incomplete. Over the past 3 months, we've done 5 major releases in our alpha, with more than 300 individual ticket items. During that time we also gradually grew out invite-only users from 100 on 4/1 to a bit more than 3500 now. We'll open the site up for general use in the next couple of weeks and call that "beta" as it's still very much a work in process, but it will be generally available.
Any suggestions as to other naming structures?
The private beta is always a good thing. You can get your product out to a limited audience and still get good feedback. On the web, there is very little barrier to entry, so many sites are being launched with one or two core features. What I was trying to say in the post, and obviously was not clear, is that once you go public, just drop the alpha/beta tag and say it is "live".
On the web, unless it is an invite-only site, why do we need to tag it as beta? I am sure most of the early adopter crowd is willing to accept some issues for a newly launched service.
So, having something like "beta" indicates that it's still a work in progress.
I do also agree though that there should be a statute of limitations on "beta" and that Gmail has more than outworn its beta label.
however, now let me defend "alpha"
I was part of an "alpha" launch this week – and it feels really good to start getting user feedback that early in the cycle. I work at Me.dium and we launched our social search alpha (http://me.dium/search) on Wednesday night, letting our users know that we’d "worked fast to get our product out there for you to play with, test, discuss, and let us know what you’d like to see next" (http://tinyurl.com/59a7vk).
We’ve had over 100,000 people try our stuff since we launched, and the feedback has been terrific – really helping us shape the "beta" roadmap and giving us plenty of tactical to-dos to bang out before then. More grandly, our users are now part of the process. They have a voice. We'll be able to build stuff they directly request. And – if we keep listening and understanding – we hope we'll get to a point where we're a step ahead: releasing new things we know our users will really love. Users help us, we please them. It might be terribly "enterprise software" of me to say it, but that sounds like a win-win.
Speaking of "enterprise software", I'm assuming that's where you are coming from with lines like: "in a product development company, you tend to send out beta software to your most prized customers." (Apologies if the assumption is incorrect). You then go on to say: "This gives the customer the chance to see what is coming". I'd say the reason for beta programs in enterprise software is a little more calculated than that... If you can get friendly customers to deploy a beta, this means you have canned case-studies to point at when you go GA. There's no better sales technique than letting customer A know that rival customer B is already using your stuff and gaining an advantage. If B was part of your beta program, you've probably just reduced your sales cycle to A by about 6 months. It's just good (sneaky?) business.
Either way, I'm very happy we're playing in the rapid release-feedback-release world of "alphas". It's also a lot of fun :) In short: bang it out, see what people think, iterate, and repeat.
The most famous company i remember bringing beta in to the web was of course Google. but funny enough the first to popularize Alpha were they guys at mozilla with Firefox.
Then you also forgot how Microsoft now handles middle term on the stage from Alpha to Beta that is now used as Technical Preview.
Live Mesh is a Technical preview because its form can change at any minute but is in no way a alpha because it is too advanced in development to be an alpha and neither a beta because the form and look it will have is not fully decided and cannot be yet promised. you also forgot a step from beta to gold: Releace candidate.
Release Candidate is for a now finished product in a form factor that needs to be previewed in order to be sure of going gold but that beyond certain touches and tweak will be exactly the same when it goes gold.
So, i think those are the most interesting terms that the web 2.0 brought into the mainstream
Back when I were a lad, an alpha release was certainly ready for public testing, and probably broadly correlates with what you now call beta. Beta, or field test, was undertaken when the product had been completed to the best of human ability, Bugs in beta test were not expected (though inevitably found). In other words, beta encompassed all three of the bottom boxes in your chart.
So yes, one can expect that inflation will cause even more boxes to be added to your chart. But like the other recent additions, they probably won'r be called "gamma."
Still, you're right, Gmail shouldn't be beta anymore, and SocialMedian is bug free (as far as I can see), and should already be in Beta, it's also a very nice product ;)
http://blog.pluribo.com/2008/06/29/say-no-to-th...
The basic point was that people who use the "beta" label don't really understand what software-as-a-service means!
Likewise journalists (pre-blogger days) as well were under the same restrictions. The last thing back then was that you wanted a buggy version colouring people's views of your product. Of course the software/sites were also paid for by subscrptions or usage so the users were really getting something for free.
Now it seems just an easy way to get a load of free testing and you mainly see freemium sites in this category. Plus of course a great way of getting a quick 100 users into the database by offering "limited" beta accounts.
A marketing consultant suggested that I apply a beta label to my HearWhere.com site so that visitors will know that it is still in development, and that feedback is welcome.
However, I think that
1) anybody who knows what beta means realizes that in the web-world nothing is ever complete
2) for anybody else who doesn't know what it means doesn't care - they aren't going to think "oh, it is in beta, I'll come back to see what has changed".
3) beta makes sense when you are dealing with native applications that have repercussions on the system as a whole, but that isn't necessary with the web.
But isn't this also a liability issue? If a service is perpetually in 'beta', then the disclaimers for possible lost or corrupted information hold up better in court, since the service provider can claim a given user is more likely to be 'aware' that a service could encounter bugs (e.g. with 'beta' stuck on the logo on each page).
An example that comes it mind is Trillian Astra, which is in private alpha because features are still being added. Once the featureset is solidified, it will move to Beta. Once bug testing in Beta is completed, it will move to a public release.