I'm not even saying they should change them, but for me, there are a few feature changes that would make it way easier for me to use (and therefore a rapid fan and repeat customer).
It's one thing to complain that Basecamp is too drag-and-drop. But not having detailed schedule management features is practically Basecamp's calling card. He's using it wrong.
Basecamp is awesome when your workflow is "generate some kind of file, post it, talk about it, repeat". A surprising number of projects boil down to that. For that workflow, Basecamp is the best thing I've used. For any other workflow, look elsewhere.
I thought Basecamp is also designed to solve the distributed project management problem in a simple, easy way. I like the appoach, but the Deadline/Time tracking doesn't work for me in it's current implementation.
Seriously. Not kidding. You want a different app. Basecamp don't want you. There's nothing wrong with that. Or you.
Their actual customers are service practices (designers, copywriters, event planners, marketing organizations, business consultants) whose workflow consists of "1. discuss some kind of file, 2. create some kind of file, 3. upload file, 4. discuss file, 5. goto 1". For that workflow, it's great.
It does not work for projects with multiple independently timeboxed steps, or for projects that have complex dependencies. The dependencies need to work like "step 2 depends on file from step 1; step 3 depends on files from step 2".
So Basecamp does not try to solve everything, it's just the best thing after e-mail.
But for task management, it still isn't better than a shared Google Spreadsheet, which is weird to me.
"No time estimates" - if you upgrade to the next highest version of Basecamp you will get time tracking functionality for all tasks, which would work just fine for basic time estimation. If you don't want to upgrade to the next highest version, then just put your estimate for the time a task will take into the comments for that task. The idea is to figure out a way to make this simple system work for you, and that's why they can cast such a wide net to so many different types of industries without catering to them specifically.
"Tasks have no correlation to Milestones" - this is not true, but probably mentioned because of the OP's brief interaction with the app. When you create a todo list you can choose which milestone it relates to. You can even set it to complete that milestone when all the todos have been checked off.
"Too much Drag & Drop" - this point is really pretty nit-picky and more a matter of personal preference. I don't claim to be a UI expert, but I could make equally compelling arguments as to why the drag and drop functionality they have in place would be better than up and down arrows.
"Why 6 categories of messages?" - This one is just laziness on the OP's part. Right at the bottom of the messages dropdown menu you will notice an "-- Add New Category --" option which allows you to customize categories. You can also globally customize categories so that they show up for other projects too.
I'm really surprised this post hit the front page of HN. It iss a cursory glance at a piece of software and a seemingly offhand damning of it. I wouldn't doubt that a few of these up votes came from people just reading the headline and having a general feeling of bitterness towards 37signals, for whatever reason. There's been a spate of anti-37signals posts over at reddit.com/r/web_design too so it wouldn't surprise me.
Basecamp is great software because it's extremely simple. They purposely keep the feature-set small and are very opinionated about how they do things. The result is a very solid tool that isn't bloated and can fit many different niches because of its accessibility. Before they came around you were stuck with overly complicated all-things-to-all-people solutions. Kind of like Rails and Apple products...
So 37S ends up being phenomenal marketers that have a service many people want to buy, but more people buy it than can get real value out of it because it wasn't built for their needs in the first place.
Certainly 37S is not ultimately responsible for people continuing to pay for something that doesn't work, but they do a pretty good sales job via word-of-mouth evangelism, even when it's a poor fit for an org's needs. Maybe too good?
Entrepreneurial companies, by their very nature, change. It's hard for me to fathom why a startup would build their software development process around a tool optimized for inflexibility, no matter how enticing, when that might inhibit their ability to fix something broken about their business later.