АБРАКАДАБРА (Тоже самое но в читаемом виде)|
Predstaviteli odnogo iz krupnix reklamnix agentstv Clix Marketing, zanimau6iesa menedjmentom PPC kampaniy, opublikovali otkritoe pis'mo k injeneram Google s pros'boy zamedlit' razrabotku proekta Google AdWords.
Skorost' razrabotki razli4nix funkciy programmi, kotoruu Google iskusstvenno uveli4ivaet dla povi6enia konkurentosposobnosti AdWords, privodit k sokra6eniu fazi testirovania, a eto, v svou o4ered', snijaet ka4estvo prilojeniy.
V odnom iz poslednix interv'u s Erikom 6midtom, CEO Google, podobnaa politika v otno6enii sokra6ennogo vremeni razrabotki, kotoraa pozvolaet vixodit' v svet «sirim» programmnim produktam bila ob&asnena «preimu6estvami, kotorie daet skorost'».
Krome togo, 4astie izmenenia interfeysa snijaut ego uzabiliti, tak kak pol'zovatelam prixoditsa zanovo privikat' k novomu postroeniu menu.
«Google i tak naxoditsa namnogo vperedi Yahoo i Microsoft, sistema upravlenia kampaniami v AdWords operejaet analogi, kak minimum, na dva i bol'6e pokolenia. Vozmojno, vi mojete sebe pozvolit' snizit' temp, prejde 4em polu4ite reputaciu proizvoditeley «sirix programm»»,- pi6etsa v pis'me.
An Open Letter to Google Engineering: Please Slow Down a Little
We really do love your software. And we appreciate the fact that you introduce valuable new features so frequently. But please: slow down a little, and spend a bit more time on bug testing.
In the time-honored model of software development (call it "pre-Web 1.0"), teams of programmers worked for years to craft scores of subroutines, knit them together into a megalithic "major release," and then test and re-test the application on a variety of software and hardware platforms. The application would often be "pre-released" to internal and external teams of alpha and beta testers who would run the software under an even wider variety of conditions.
All this methodical testing slowly but surely eliminated major software bugs until the "release" could be dubbed "Golden Master." Only then would it be released to the buying public, hopefully with only minor bugs remaining. Thus Word 2.0 begat Word 3.0, etc.
That model, though still practiced widely for PC- and server-based applications, seems almost anachronistic in today's environment of high-velocity incrementally-released Ajax-based web applications. New features – minor and major – appear overnight, often with little or no warning, explanation or documentation. And all too often, with minor and major bugs.
As an agency that manages client pay-per-click (PPC) advertising campaigns, we at Clix Marketing are elbow-deep in Google AdWords software – all day, every day. It's not unusual for us to wake up in the morning and find old features completely replaced by new ones. Almost always the new features provide improved functionality, or eliminate time-consuming steps from a process.
Occasionally, it's painfully obvious that the new feature hasn't been thoroughly tested. Recent examples come to mind: a new graphing feature simply failed to work (and displayed only flat-line graphs), and it took Google a week to fix it. AdWords Editor is up to its fifth release, an age when most software has attained stability – and yet it still crashes frequently.
The phenomenon isn't restricted to Google's AdWords software. Last week the phone development community was up in arms over their contention that Google's Android platform for phone application development "…has major bug issues."
A recent New York Times article explains that Google's breakneck software development speed is a concerted, encouraged effort to stay out ahead of serious competitors like Microsoft. "Velocity matters," says Google CEO Eric Schmidt. Google product road maps look ahead only four or five months at most.
But here's the thing, Google: you're already waaaay out ahead – for example, your Google AdWords campaign management software is two generations or more ahead of Yahoo and Microsoft. So you can afford to slow down a bit and test more rigorously before even minor software releases. You're starting to get a reputation for developing buggy software – one that's not deserved for the most part – and now is the time to nip the problem in the bud before the perception starts to snowball.
Thanks for listening.