> There are many powerful, competent, and well-funded entities with a vested interest in improving the web app experience (including Google). I wouldn't claim they're doomed from the start.

I've been using gmail for almost 10 years. I liked it from the start. But it's getting worse, not better. It's extremely slow compared to what I'm used to with Exchange and Outlook, and navigating around is more of a chore. How do I, for example, see who's in on an email list? I can't just right-click on the name of the list and see the members? (And don't even get me started on the new compose experience mess.)

If Gmail was being continuously improved, I'd have a reason to be optimistic about it. I like being given free storage, but as far as usability goes, I have fonder memories of Eudora.

Gmail UI sucks in part because they are going for the masses, not the power user. It isn't HTML5's fault.

They are going for the masses, but my mother complains to me about it (the compose experience). She's 60.

Don't get me started on the emails she loses in the Promotions tab.

in case you don't know it, you can get rid of the tabs for her, so everything goes to one inbox again.

I'm just saying this because, to _remove_ a tab you have to click on the "+". Awesome intuitive design.

A lot of it is due to the limitations of HTML. For example, HTML is why Gmail paginates about as well as PHP app from 1998, instead of allowing you to just scroll through all of them like a native client.

Nonsense. FastMail's webmail allows you to scroll continuously through (or jump to any section of) your mailbox, even when they have 100,000+ messages in them. (Disclaimer: I work for FastMail). This is not an inherent limitation of the underlying technology, just one of implementation.

> For example, HTML is why Gmail paginates about as well as PHP app from 1998, instead of allowing you to just scroll through all of them like a native client.

No, its not. HTML + JS, via AJAX, supports infinite scroll with a finite amount loaded at any given time. IIRC, Google has used that on Google Image Search and some other properties.

There may be web platform associated performance or other considerations behind pagination, but its not a fundamental can't-do limit of the technology.

Pagination does have it's advantages, namely: - Linking - Easier to grok where you are in the result set ( 3 out of 5 instead of the scroll bar that changes height)

Yeah, it works great in search results, much easier to understand for the average user...

Page 1: 1-20 of about 26... cool not too many to scan through

Page 2: 21-40 of about 46... oh

Page 3: 41-60 of about 67... errr

Page 4: 61-80 of about 87... seriously?

That could definitely be part of it. Outlook isn't dumbed down because it's not primarily aimed at home users. Maybe Gmail makes sense for few people in the office the way Outlook makes sense for few people at home. That doesn't change the fact that the bandwidth/latency profiles are different going to the server room in the building than going out to the internet. But I might see a big improvement over what Gmail is now if Google was being designed for business users.

What features would you add to Gmail to make it more useful for business users?

What UI fixes would you make?

How do I get started writing plugins for gmail to add the above features and fixes?

For starters Google could of kept compatibility with the activesync protocol instead of making calender sync with outlook the horrific monster it is today. Actions like this are just one of the multiple reasons I tell people to not invest invest Google business products.

Amen. Gmail has _lost_ features over the years, as opposed to gain any.

Perhaps than they should give the masses their interface and give us a standards-based/compliant JSON interface. We're headed that way already.

That's IMAP, CalDAV, SMTP isn't it?

Where does JSON even need to come into it?

Because I could then build the entire interface client-side in a browser?

Don't get me wrong, I'm not a webdev guy. I have no front end/fullstack experience. I'm a longterm sysadmin/DevOps person. But if you're a power user, you want control over your mail interface/workflow. In that case, all your mailserver should be doing is accepting email for you, storing and indexing it, and serving it via API to clients you're using. I like IMAP, but it doesn't easily support some Gmail conventions (multiple labels per message).

IMAP and SMTP could easily be condensed into an XML/JSON API that could be done over HTTPS; I'm not familiar enough with CalDAV to say that though.

Which JSON-based email standard would that be, exactly?

I believe ops point was that any stable api would be preferable to: "Yeah, there's an api, but we refuse to document it, and we'll randomly depricate stuff if you try to use it to build something that isn't gmail". The team behind gmail is probably one of the best qualified to hammer out a working api for email of json (what we have + a bit of what we want + stability and versioning). No reason why they couldn't publish that as an RFC and let people implement a front end for dbmail or whatnot that spoke the same api.

Except for competitive reasons and opportunity cost.

Thank you. This is what I was getting at. I hope we get closer to this with https://www.mailpile.is/

Not a standard yet, but this just made it to HackerNews: https://nvlope.com/

"nvlope gives you full control over your mail with a full-blown JSON API. You can build an email client with just 3 API calls - threading included!"

http://developer.nvlope.com/

He said including Google, not including Gmail.

Gmail should not be indicative of whether or not web apps can match desktop apps. It should be indicative of whether or not the engineers behind Gmail have focused on making it work as well as a desktop app (they have not).

They've spent a lot more time making Gmail Offline work well. Maybe sometime in the future they'll spend time optimizing; but it's probably not a priority

>> It's extremely slow compared to what I'm used to with Exchange and Outlook, and navigating around is more of a chore.

Its performance is fine from what I see every day. Plus it integrates with my phone and my home Linux desktop via the browser very nicely. Navigating is different but not worse than Office by any stretch, just my opinion.

Some of us don't like any of the current alternatives.

In my opinion Gmail beats Outlook for mail, but Google apps are not really enterprise-ready for calendar and docs. Editing a document offline is not a feature that should just now be becoming sort of usable, it should have been fundamental from the beginning. Confusion between Google Apps and personal Google logins is common. Managing a Google calendar is painful, and basically has not improved at all in years.

I have no desire to go back to using Office for everything, but I have very little confidence that Google is actually going to spend much effort in getting things right for these use cases either.

> I wouldn't claim they're doomed from the start.

They are set back and certainly extremely limited from the start. Native means you have almost unlimited capabilities for interacting with the user and their system. A browser application by definition is much more limited. The experience is always improving but it's to get the point that native applications have always had -- and they still have a long way to go. And the nature of browser based applications might mean that it can never happen. The browser is layer between the solution and the user and it will always interfere in some way.