The Haiku Devs should put a larger focus to the web browser

1

I know it sounds strange, but I think it’s best.
I’m not demanding anything, or saying it’s something you need to do right now.
There’s no OS quite like this one on the market, so I think you have plenty and plenty of time. Whenver you feel like it.

This OS runs really fast and boots fast, so it’s already great for netbooks.
Also, netbooks are called “netbooks” because a large focus on them is for the net.

I’m basically saying that if a much larger focus was put on the web browser, then the OS itself would already be very usable.

Just look at it now. The HTML5 support is already decent. So, if I am having trouble finding a program for Haiku, I can find it in the web browser.
Like, I couldn’t get any Haiku office suites to load, so I just use Google Docs. It’s a great office suite that runs in the Haiku web browser (Web Positive).

I’m just saying that if Haiku could support what modern browsers support and the great speed it does, you’ve already sold Haiku as a great netbook OS.

Then maybe you could improve upon it to get it to be a great desktop OS.

This is just a suggestion. I bought Haiku R1 and I think, with just a bit of additions to the web browser, it could actually be a main OS for me. I mostly just want Flash and WebGL. But if somehow you managed to get Java, Silverlight, and all that stuff over, that would be pretty impressive. But I don’t think Java and Siverlight are much requirements.

I also think a spell-checker would be useful as well, and eventually extension support.
As well as making it load pages a bit faster and being able to freely move the tabs around.
Other things like that.

These are just my 2 cents. It’s summer now, and I’m working on my own projects.
You don’t need to take my suggestion if you don’t want, but I’m just sayin’.
This already has the potential to be one of the best netbook OS’s. If you can get all of the stuff supported, WebGL, Flash, Java, Siverlight, etc, it could even be better than Chrome OS for netbooking. It certaintly has the speed.

I know the devs probably don’t want their OS to only be a netbook OS.
But if you want it to be a good desktop OS, you need to at least get it recognized first.
I’m just saying that if you get it recognized as a great netbook OS, more people will start using it, and then you could get a crowed.

I honestly have no clue how hard it would be to support the things I listed, but hopefully one of you can, or at least know how to find somebody that can.

I’m posting this from Haiku. I think Haiku has a lot of potential, if the project is focused in the right direction.

You may be still waiting for better financing, and that’s okay. But whenever you get it, I just suggest putting a large focus into the web browser. It’s already looking pretty good.

The web browser is just an easy way to sell a fast OS right off the bat.
Or if you do not plan to sell it, at least get it recognized.

Like I’ve said, even if you don’t want Haiku to be recognized as a netbook OS and want it to be something more or different, you want to get it recognized first. And this would be a good place to do that. I don’t think Haiku is far from being recognized at all.

2

Hey nitt,

Thanks for your suggestions. For what it is worth, I completely agree. I am the one who originally ported WebKit to Haiku (what we built WebPositive on) and I’m the closest thing we have at the moment to someone who works on WebPositive. Unfortunately I’ve been busy doing work and other projects and haven’t worked on Haiku in a long time.

But we definitely need to greatly improve our browser. I don’t think it even has that much to do with being a good netbook OS. Whether Haiku runs on a netbook, a desktop computer, a laptop, or even a tablet someday, we have to have a good, modern and fast browser.

Honestly a big part of why I don’t use Haiku that much myself (even though I am a Haiku developer) is that WebPositive is not good enough. Of course that is as much my fault as anyone’s.

Fortunately my interest in Haiku and improving the web browsing experience has been rekindled lately. There have been some interesting developments in the web browser market lately, such as Google breaking off from WebKit and making their own browser back-end called Blink (which at the moment is just a copy of WebKit), which they will use in Chrome. I intend to do research into whether Blink would be a better choice for Haiku than WebPositive, but initial looks show it would probably be a large amount of work to get it going.

The main issue here is that the web is a complicated platform, and it is getting more complicated by the day. There are hundreds or maybe thousands of people who work on each major browser platform (Chrome, Safari, Firefox, Opera, Internet Explorer) and Haiku is already spread quite thin on number of developers. I have plenty of other projects I could help on in Haiku, but since no one else seems to have the time or motivation at the moment to improve the browser, I’m as good as anyone. But the reality is a lone developer can only get so much done in any given amount of time, so development will likely be slow, especially because I have a day job and other commitments. Of course slow development on the browser is better than NO development!

To conclude, yes we need to improve the browser, and I will try to work on that. Unfortunately even just full HTML5 support is a big project, so things like Flash or Java support are even more unlikely to come anytime soon (especially Flash, since it is a closed platform.)

3

leavengood: Most agree with you. But there is one case that completely denies that the browser has to develop to system staff of people. This is OWB for MorphOS system created by one man: Fabien Coeurjoly. Previously use Oddyseys engine, now “pure” WebKit. Please take a look at the link what this browser can: http://fabportnawak.free.fr/owb/owb-morphos-1.20.readme
I do not know whether this man is a miracle worker or not, but practically alone develop it - WebKit update and add new features.

4

Looks interesting.

One thing I’m a little envious of just from the readme: support for Rexx scripting. For example, a SELECTEDTEXT command - I had an application where I could have used that. I’ve tried to work WebPositive with scripting messages, but it seemed (at the time anyway) to support only whatever minimal access it inherits from graphic primitives. And I am not really sure it makes sense to spend a lot of time adding to that, as my hunch is that BeOS application kit scripting may not really be all that viable, compared to the ARexx thing.

5

The developer of Odyssey Web Browser is definitely a good and skilled developer from what I can see, but he has been working on it since February 2009 (actually that was the first public release in the release notes, he was probably working on it before then.) So it is not like he did all that in the last 6 months or something. Of course our WebKit port is a bit older than that and WebPositive was originally created by Stephan Aßmus in February 2010. It is definitely inspiring to see what a lone developer can do if they consistently work on a browser project.

Our issue is that the developers who can work on our WebKit port and WebPositive (Stephan, Alexandre Deckner and myself) also spend time working on Haiku itself and of course we all have the usual “real-life” concerns such as work, family, hobbies which we spend time on. We all also tend to take time off from Haiku development for fairly long periods.

Also working on WebKit is quite difficult: the code base is absolutely gigantic (probably as big as Haiku itself, or bigger) and even operations in the source control tool Git are really slow (particularly getting the status of changes) because there are so many files in the repository. Small changes in key WebKit header files can cause almost the whole project to require rebuilding, which takes hours on a slower machine. The code is also so large that it is essentially impossible to compile the whole thing with debugging symbols, or more importantly to load that debugging-enabled build into a debugger. Since it is such a large and foreign code base, finding bugs is quite difficult, and the issue with the debug version just mentioned doesn’t help the matter.

Also since many different companies employ hundreds of people to work on WebKit, the project changes at a very, very fast pace, and keeping up-to-date is almost a full-time job. Their abstraction for multiple platforms is quite poor, and involves quite a few #ifdefs and other unsavory techniques which can cause more changes in platform related code than might happen in better abstracted systems. As a result our port of WebKit was removed from their repo because we were not maintaining it and they were sick of “fixing” our port whenever they decided to yet again move things around.

So to add insult to injury, the Haiku port we have now isn’t official at all, and only lives in a branch on GitHub at the moment. Updating is then all the more difficult because as far as WebKit is concerned, the Haiku port doesn’t exist. Updating our port involves updating WebKit and then fixing everything that was broken before even getting a chance to try to add new features or fix bugs.

And that is just the core HTML, CSS and JavaScript part of WebPositive! Beyond that we have the browser itself which is where things like bookmarks, history, tabs, windows, autocomplete, etc are implemented. Ignoring WebKit, improving the browser itself is also a lot of work. And certain features, like a password manager, require other things to be implemented in Haiku before they can be implemented.

Anyhow, enough excuses I guess, but those are the reasons why development of the browser has been so slow, or downright stagnent.

I suspect I’ve written all the above more to make myself feel better than to inform those of you reading this, but oh well :wink:

6

I don’t know much about Rexx, but I think we could support most of the operations in the OWB README with a BeOS/Haiku scripting system in WebPositive. It is just a matter of implementing it, and this is one of those things which would be fairly fun for a developer like myself, so I’m sure I’ll do it at some point. But like a lot of things in Haiku there are more important things to do first on the browser before doing fun stuff (though I certainly skip to the fun stuff sometimes when working on Haiku to keep myself motivated.)

7

Cool, well, it’s a powerful tool. My application was a web service client that was pretty self sufficient but needed a token for authentication that had to be obtained via interactive browser session. So the application could invoke the browser to get that started, but I had to copy and paste the result from the browser to the client. I guess it isn’t a big deal, but if the application could have driven the process the rest of the way it would have been easier and more robust. I’m sure there are more interesting applications for this.

8

Excuse me, but we can not focus on Arexx because it is the least important in this topic. Thx guys.

leavengood: I did not write that he did it in six months or a year - let it be, that it took him four years but he did it practically alone. Approx. we have some younger browser but a little more people on the working (About WebPositive). Effect? The application, which today is one of the most important in any system in our lame. But with us is to protect the memory multiprocessing, virtual memory - this does not have MorphOS. Meanwhile, using W+ to view you get neurosis, an application like a long time to wonder any longer upload the page - and it is the fault of the browser and not the network stack, because the comparison can launch any other browser for Qt on Haiku is “thoughtfulness” there does not exist.
I understand how large project is the same WebKit, but in the end you have to point out that anyone who installs Haiku certainly will use from WebPositive - and, unfortunately, but the action that he see can the negative impact on the perception of the whole system. Simply, W+ should have priority in the development of a system application.