TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

Enterprise Software Doesn't Have to Suck

26 点作者 jsatok超过 14 年前

6 条评论

Locke1689超过 14 年前
I think some people who write about enterprise software either haven't worked on really big enterprise software or don't realize how it came to be bad. The problem isn't that suddenly your product started sucking -- mostly it's that inefficiencies or poor design decisions made 3 years ago suddenly start biting you in the ass.<p>In other words, software ages like people -- year to year there's not a huge difference, but 10 years later you look back and say "what the f<i></i>k happened?!"<p>Now, some software just clearly wasn't designed with usability in mind. Much more, however, was probably designed with a limited goal that was forced to quickly adapt to more complex situations. Startups and web companies have it relatively easy -- you can simply change your code on your server and suddenly everyone's up to date. Other companies have to deal with backwards compatibility and standards and shipping and all the pains of large projects in general. The people who come in saying "Well the people in charge are just idiots. If <i>I</i> were in charge we'd have SCRUM and new languages and modern development practices and everything would fix itself immediately!" really don't know what they're talking about.<p>I think that the article has some good points (not getting involved enough with the actual users is a big one), but they're really too young to say "we have found the Holy Grail!" Removing features, especially, strikes me as something which sounds like a good idea but can really hurt you as a business. When your customer is a 300 person mid-size corporation you can easily say, "well, we had to make some changes here and that feature had to go." When your customer is a retail giant employing 300,000 people all over the country, you think very, very carefully about removing that feature.
评论 #1719726 未加载
评论 #1718637 未加载
Poiesis超过 14 年前
The article gets it right in defining the classic problem: buyers are not generally users. And when they are, they're not qualified to make informed decisions about design.<p>Where the article gets it wrong is the solution. At least in my limited experience, enterprises are slow to change things if they "work", however painful. They also do not always entertain SAAS for a variety of reasons, two of which are probably inertia and a different sales process than they're used to.<p>Couple that with IT departments that block "productivity" or "personal storage" sites and you can see where alternatives can't even work their way in slowly.<p>I think the solution described is the best chance we have of getting better enterprise apps, but it's by no means going to be easy. It does seem that over time things will get better. It's already been far, far too long. I hate using broken tools.
评论 #1717957 未加载
jnoller超过 14 年前
I completely agree - so much midtier and enterprise software is a complete mess because of these things. People have to be willing to listen to customers, but say no when it serves the greater good. You must kill ruthlessly - but you must have the metrics to know what to kill, and you have to examine those metrics as part of the process.<p>Feature bloat is especially painful, for the longest time "enterprise" has been synonymous with "a ton of mostly unused features and confusing UI". Don't make your product into something which resembles a pile of tools in the crap-drawer in your kitchen. Say no - kill things, make it clean, organized and usable.<p>Your users will thank you. On the other hand, the guys who make more money than you every year "consulting" on how to "properly use and setup" you bloated confusing product will hate you.
评论 #1717761 未加载
Flemlord超过 14 年前
Had me right until the end: "Solution: Kill features. Often."<p>In most enterprise environments, that's not realistic. Unless you've really screwed your design process, features are in there because clients need them. Even refactoring a feature into something that makes it a better product for 99% of your users will greatly annoy the one user who needs the feature and is used to having it implemented in a certain way.<p>It's much better to negotiate with the users on the front-end for a better product or avoid overly-demanding users that will hurt your product. Because once the feature is in there, it's there to stay.
评论 #1717576 未加载
petrilli超过 14 年前
The problem, as I've seen in a few big organizations, is that everyone thinks they need an "all singing, all dancing" piece of software (like SAP or Siebel, for example), which has 18-quadrillion knobs to adjust. The problem is, nobody has the slightest clue how to adjust those knobs to fit your business, but the sales person convinces the customer that somehow, magically, it will be a perfect fit.<p>What ends up happening? The customer has to change their business to match the "default" behavior of the infinitely flexible software, because nobody can overcome it's intrinsic infinite complexity.
jaygoldman超过 14 年前
Dan (jman73) has posted a follow-up: 4 Steps to Becoming a Killer. Comments and discussion welcome! Post: <a href="http://rypp.ly/9XKdvm" rel="nofollow">http://rypp.ly/9XKdvm</a> HN thread: <a href="http://news.ycombinator.com/item?id=1744613" rel="nofollow">http://news.ycombinator.com/item?id=1744613</a>