What a massive failure. I wonder if they realized they weren't going to hit the deadline so they decided to do a quick implementation with the tablet app. The old rumors that Gmail is a massive spaghetti mess must be true. Let's just hope they're actively rewriting it and don't think this is good enough.
One thing we learned early on was that making any existing system (especially one as big as Gmail) work offline had some very difficult, fundamental issues. The biggest of these is that everything that normally happens server-side must also be able to happen client-side when there's no network connection.
This is a problem not only because there is a lot of server-side complexity in Gmail, but also because it changes all the time. New features are added, old behavior is changed, and if the offline code path isn't changed along with it, things have a tendency to break horribly. The only solutions to this involve having every Gmail engineer modify the offline code path whenever they touch the online, which is a large burden that we weren't willing to lay on them.
As a tablet user, I actually like this; when I got a honeycomb tablet, I realized I had come to miss the split-screen view after several years of using Gmail...not that I want it to turn in gOutlook, but there are definite advantages.
On the negative side, it's very weak that there are no formatting controls in this version. I miss them on the tablet side already, which means that messages sent from that device necessarily have a different signature from those sent from my desktop. It's not necessary or even desirable for mail recipients to know I was using one device or another. Some people like to signal that they are on the go with 'sent from my phone' or the like, but I would rather have some basic editing functionality, and this would not be hard to add. On a tablet it's annoying, on the desktop or a laptop where I might need to work offline it's unforgivable. On the up side, this is a good starting point for building an editing/configuration interface that could find its way back towards the tablet space. With wide-format screens the norm nowadays, I was struck by how much more pleasant the panel experience was on my desktop monitor and could see myself switching to this from the page+widgets approach of existing gMail, which is beginning to feel very long in the tooth.
This goes double for Google Docs, although it's slightly off-topic. I like working on my tablet a great deal and find I can type surprisingly fast even with the on-screen keyboard as opposed to an external one. But Google Docs on a tablet is so unusable that every manufacturer ships with some open-source office suite to make up for the deficiency. since Android does not have the same mind/market share as iOS/iPad, Google needs to offer compelling software alternatives. In many respects it already does so; but productivity tools are noticeable by their absence or abridgement. A minimally-capable version of Google Docs made sense on a smartphone, but on a full-size tablet it's absurdly self-defeating, and inimical to a corporate or academic environment. I would pay for a 'power user' tablet version of core apps like gMail and gDocs. The fact that there is no easy way to leverage forms into an Android clipboard/data-entry interface is a major missed opportunity.
This really needs to see some movement in tandem with the upcoming Ice Cream Sandwich launch. I know both Docs and Android are outside of your remit, but hope you could circulate these concerns in the cafeteria, so to speak. I mention this here because *@gmail.com is increasingly acceptable as a professional address, but this will only last so long as the tools cater to the needs of professionals.
I think the difficulty is that none of the technologies that you are using were designed for building software applications, but for showing text and images on a screen.
In other words, there is a perfectly fine solution for offline G-mail, it's called "e-mail client".
Actually the problem may be that OS user interfaces, for all their native functionality, have not kept pace with browsers at showing text and images on a screen. Hypertext is an innovation that is hard to ignore.
Native UI is vastly more powerful compared to HTML and CSS, which were designed to display web pages, not widgets and controls.
In fact I don't understand what kind of experience you have to say something like that. If you were a web designer you would know that the web UI is a painful to use hack.
If you were a mobile/desktop developmer you would know that both types of UIs are better at showing pretty much anything on a screen - including a browser engine.
I can't comment on unannounced engineering efforts, and I certainly didn't mean to imply anything about them.
All I can really say is that if they wanted the primary Gmail client to support offline, then it's my personal opinion that the only tenable way to do that is by having a single code path. Whether there are plans to actually do this I don't know and couldn't talk about if I did.
Maybe they just want a solution that people can use RIGHT NOW. What's wrong with that?
I'm thankful for Google providing this solution though I'm hoping for a better one in the future.
Calling it a massive failure is just mean-spirited.
You call yourself an "aspiring entrepreneur" but I don't think you understand what it means to be one and actually build stuff. If you do, you wouldn't be making comments like that.
I'm a programmer and I've received much harsher words about stuff I've written. Maybe I'm hypocritical, but I feel less of a need to bite my tongue with Google than with a HNer's weekend project. The fact is that they've been promising this feature since they discontinued Gears over a year and a half ago and to then fail in this way is both disappointing and insulting to me. It's insulting because I'm passionate about the web and I know that web applications can handle offline just as well as native applications, and their slowness to add support to their most important products contradicts that belief. And it's insulting because they've released something so that they can say they have Offline Gmail when this barely counts as that. Having separate apps, with separate UIs, and separate ways of launching, for an online and an offline version is not a customer-centric decision; it's a face-saving PR decision.
"It's insulting because I'm passionate about the web and I know that web applications can handle offline just as well as native applications, and their slowness to add support to their most important products contradicts that belief. And it's insulting because they've released something so that they can say they have Offline Gmail when this barely counts as that."
Insulting? I don't get it. What are you entitled to? That is such an arrogant statement. Can I ask what have you done to the web community since you are very passionate about it?
Showing passion is not about putting other people who actually build things down.
You have to realize that Google may be a big entity but it is still a group of individuals.
You have read the comment from the Offline team and they have a good reason why they can't do it right now.
Gmail is probably Google's biggest product after search. It's also been around for a long time, so you do have to keep that in mind.
I would presume the Gmail team are still working on offline, but Google prefers launch + iterate, rather than waiting for things to reach a "finished" state. This was the quickest way to get people to an offline Gmail product.
DISCLAIMER: I'm an intern at Google, but have no information about these products apart from what is publicly known.
This comment is completely out of line. Trashing something pretty decent on its launch day is not a sign of elevated or discerning taste, it's a sign of someone who does not understand how hard it is to ship something.
Having some kind of offline Gmail (and Docs and Calendar!) is indeed a VERY big win, even if it doesn't have every conceivable feature on day one. It makes Chromebooks that much more appealing.
Ok, the UI and launcher are different. That's not "massive failure", that's minimum viable product. Yeah, even Google does that.
Since you develop for the Chrome app store, did that particular comment advance your cause? Are there particulars you could point them toward, or identify things they should keep in the total re-write you hope for?
I don't hope for or care for a total re-write; that's up to them. But they should be able to cache all resources with AppCache. They should be able to cache recent messages with LocalStorage. These are not new technologies; they've been around for a couple of years at least. Here's Google discussing their use of AppCache in Mobile Gmail well over 2 years ago.[1]
Yes these technologies have been around for several years. Unfortunately none of them are stable/reliable/easy to reason about.
WebSQL (SQLite) has been ignored by Mozilla, on the stated basis that an independent implementation of WebSQL would be too difficult to build. They advocated that a database designed by committee would be better for the web than the most widely deployed database in the world. They deprecated WebSQL before IndexedDB was given a chance to prove itself. Unfortunately they also said that anyone could build SQLite on top of IDB if they chose to. Firefox have implemented IDB on top of SQLite. IDB is at least an order of magnitude slower than SQLite.
LocalStorage is capped at approximately 10mb and there are no quota apis to adjust this. It is not likely that this will change and there has been discussion on the Public Web Apps over whether LocalStorage should also be deprecated in favor of IDB.
ApplicationCache works well but it's designed for caching code and html and it's not an alternative to IDB.
IndexedDB is high-level and "to-do list" friendly. It is not primarily focused on providing library authors with performant low-level storage primitives (BTree, KV) with which to work. The specification is exceedingly complex and so is the API. Chrome has implemented IDB on top of LevelDB which is a terrific storage engine. Exposing something like LevelDB directly would be more useful and make for a powerful, flexible API to boot. IDB has problems with inadvertently capturing application state bubbles. Indexes must be predefined and then migrated at the database layer, the application has no say over indexing. I'm not sure if IDB supports indexing object array values yet. IDB claims to support MVC yet most implementations cause writers to block readers and even readers to block other readers.
Edit:
Mozilla don't plan to support the FileSystem API.
It's not clear that all browsers will provide persistent storage guarantees. At least Mozilla have said that they may treat client-side storage as evict-able at the discretion of the user agent (as opposed to user), rather than as persistent storage. I'm not sure if the specifications have been updated to enforce some kind of guarantee on the part of user agents.
Well, there's only about 500 users with 2000 or more karma, but I'd still say that's roughly the baseline for "reputable": you've stuck around and made some contributions.
In so far as this site only seems to keep the interest people who are interested in this scene, that seems like one of the many types of folks a googler working on this would want to do right by. And, it would seem to me that his comment wouldn't exactly make a googler feel a sense of accomplishment. Maybe you're right, they might dismiss it. But there's a non-zero chance they felt a little worse, their lunch didn't taste quite as good after reading that. And if that's the case, my point to him was to ask what purpose did that serve? Because it didn't seem like a nuanced criticism.
Geez, I certainly don't want anyone to feel bad about themselves or have a bad day because of some flippant worthless comment I made. My opinion really isn't that important.
It's still August. Stuff doesn't get rushed out until the end of a quarter when OKRs are looming. At that point you can push any old thing and get away with it.
In other words, check back in a month to see even worse.
