Whither a new browser? Or two?

From a different topic, I was reading with interest, but it raised a question about the status of Iris, namely is it still beta? I ask this because the issue of new browsers seems very unclear.

It seems that some people can have it. They seem to be referred to as Shareholders. Shareholders of what, and why do they get it? It also seems that if you buy a computer from RComp, you get it, or a version of it. It might be called SomethingBrowse, but it’s clearly Iris rebadged. So it seems to be good enough to be distributed with a system.

From this post, it was due to be released “in a few months” but that was nearly eighteen months ago. It mentions something about “a support burden” but after eighteen months, surely a lot of that has been ironed out, and it could easily be released on the basis that community support is the only way to get help. Most things on here seem to be “community supported” and most things seem to have a forum or a mailing list, so what’s wrong with that approach?

I’ve seen people ask at shows/meetings about how to get a copy of Iris or FortyBrowse or whatever, only to be told to buy the computer it comes with. I’m not sure I want to buy a new computer just to get to use a browser, when my current Pi setup works well enough.

More interestingly, I saw a post on Twitter last night, asking a similar question, after someone called @RISCOSDev said THEY too had a browser for RISC OS? I assume this is OWB that used to be talked about many moons ago. So are there two new browsers, one developed by RComp, which is only distributed to their shareholders and customers, and another developed by @RISCOSDev? If so, why are two RISC OS companies developing two browsers in parallel, when people only need one to work?

It is very confusing.

It seems that some people can have it. They seem to be referred to as Shareholders. Shareholders of what, and why do they get it? It also seems that if you buy a computer from RComp, you get it, or a version of it. It might be called SomethingBrowse, but it’s clearly Iris rebadged. So it seems to be good enough to be distributed with a system

Iris is still offically in beta and has been given to the investors in the RISCOS Developments Ltd, because they have paid towards it developments. Owners of !OBrowser who brought the CD will also get it next I understand then a more general release.

A cut down version is supplied with the ARMBo(o)k laptop supplied by R-Comp. Not sure if it supplied with anything else.

OWB was the other browser that was being initially developed also via RISCOS Developments Ltd, but I think, its no longer being developed or much more slowly be worked on.

RISCOS Developments Ltd are paying for the new browsers. Andrew Rawnsley one of the directors of RISCOS Developments – owners effectively of RISCOS 5 also happens to be the owner of R-Comp who sell the ARMbok.

Hope that helps.

Thanks Chris Hughes, that sort of cleared some things up. But I have a few more questions, but please don’t feel obliged to answer.

I was under the impression that Iris was an RComp product given that it was distributed, under different guises, with their hardware. I wondered if it was a development on their browser from back in the day. Webite, was it?

I’ve done a bit of web browsing and video watching since the reply and noticed that @RISCOSDev did indeed launch both Iris and OWB with the intention of them being made freely available to all RISC OS users. Is there any reason that users of OBrowse, whatever that is/was, are scheduled to get a priority over other users? Is/was OBrowse an early version of Iris? It ties in with the other version names, like PineBrowse and 4teBrowse (not FortyBrowse as I said before).

On that, how is the bundled versions cut-down? Are there features missing? Or is it like the old ‘offline only’ browsers that used to be available? I was impressed with the accuracy of the browser when demonstrated in one of the video talks, so I look forward to it being released wider.

Is there a timescale anywhere for that to happen? Eighteen months seems like a very long time for something so vital to RISC OS to be in beta. It seems more likely that it’s about ready to go if it’s being shipped on new and existing systems. Has anything been declared in terms of actual timescale? I can find references to ‘a few weeks’ and ‘by Christmas’ (or it may have been the end of the year on the video) but they all seem to have passed already.

That last paragraph clears up some stuff, but I had no idea about the interconnectedness of the browser with those other things. Given that I didn’t know who @RISCOSDevs were, it does leave me wondering what RISC OS Open’s role with the browser is? Are they doing the beta testing? I’ve checked the latest disk image and there’s nothing on there.

Can R-Comp give some light on this questions?

Specially, if it is planed to release a new browser ( payable through !store as example) ?

And what is the expected time shedule, if so?

According to Chris Hughes, @RISCOSDevs are the ones to answer, but I have never seen them on the forum, which I find strange if they own the operating system.

According to Chris Hughes, @RISCOSDevs are the ones to answer, but I have never seen them on the forum, which I find strange if they own the operating system.

Andrew Rawnsley is a regular contributor to these forums.

Some answers:

• OBrowser is a slightly enhanced version of Otter-browser, which is a Qt-based WebKit-based browser. Otter-browser is a straight port of the Qt variant, using the RISC OS Qt port – therefore it behaves very non-RISC OS-y
• Iris is a WebKit-based browser but with a lot of work done to integrate it better into RISC OS
• OWB is a browser based on a special abstraction layer that makes porting easier, but is based on an old version of WebKit. It is said to be a lot less resource-hungry than Iris (i.e. modern WebKit)

And why does it take so long? A modern browser is probably the most complex piece of open source software you could try to port to RISC OS. A modern browser assumes an OS capable of many things that RISC OS does not provide (a beefy CPU and native multi-threading, fast I/O, accelerated graphics, a large amount of readily-available high-quality libraries, a modern font subsystem…). And porting it needs a reasonably modern compiler. So if you start a port, you start with porting a new GCC first, then port the zillions of libraries needed, integrate them as good as possible into RISC OS, work around the various problems related to assumptions about the OS…

Did I mention that it is a lot of work?

R-Comp/RISC OS Developments decided to carry out much of that work not in public. Same for the beta versions – very restrictive circulation up to now. Whether that is a good idea or not, I don’t know. Other platforms are very successful with a more open, collaborative approach. But it depends a lot on the developers doing the work – RISC OS has a long history of one-man projects.

Richard H and Stefan Huber,

Thank you for your answers. I didn’t realise that R-Comp and @RISCOSDev are the same company. That would explain how RComp are able to include versions of Iris with different names on their machines. I thought they were two separate companies, and it seemed odd that RComp were able to do that when a @RISCOSDev were paying for the browser development, as Chris Hughes said, but now you’ve explained they’re the same company, it makes a lot more sense.

Thank you for making that clearer.

R-Comp and RISC OS Developments are not the same company. But Andrew is involved with both.

Thanks Steffen, that is where my confusion will have come from.

Do we know if @RISCOSDevs will be making the browser open source? Surely, with it being on an open source OS that they own (and made open source), it makes sense to open source the browser too, especially if it’s going to be free to all. After all, the WebKit component is open source, isn’t it?

Although I notice we’ve heard nothing from Andrew Rawnsley or @RISCOSDevs about that still indeterminate date for some sort of release to general RISC OS users.

I would like to add two points:

- I think the main reason why it has not been made open for everybody is that it is still in beta and the developers probably don’t want to get a lot of bug reports as long as they have their own list of bugs to fix anyway. So releasing it now for everybody may slow down development. However if you want to try it now you can still invest into RISCOSDev to support the development of the browser (and their other projects)
- AFAIK they also work on the second browser OWB. The reason is that Iris uses webkit which will render most pages correctly but can be quite slow. OWB was meant to be probably less compatible but faster because it user a smaller engine. Just my understanding and I don’t know the current status but I think that was the initial intention.

There are free capable alternatives for RISC OS with Netsurf, Otter, Qupzilla etc though with some limitations including speed for the latter two.

I know people are eager to get their hands on Iris but ROD have been clear about order of release Investors, Beta testers, those who paid for OBrowser etc and it is technically still beta and if it was made generally available with known impacting issues what would that do for it’s reputation or indeed the amount of work feedback to what is a very small development team.

As someone who managed new product introduction and trials I know it is always a balancing act between managing expectations, wanting to release something and making sure it wasn’t killed at launch by bad feedback or indeed missed the boat because you delayed it too long.

ROD, via Andrew , have said those that purchased OBrowser will be next before more opening up and with Virtual Wakefield getting near who knows what may get launched?

Modern RISC OS capable systems can usually spin up Linux; which is a dull and boring way to run a browser.
Best of luck though; to the brave ones.

Well, I’ve had some fun today with the browsers that Doug Webb recommended.

After much reading on the forum, I added the VFP packages channel to PackMan and downloaded QupZilla-VFP. And very good it is too. Definitely not RISC OSy but works very well. I would like an icon bar icon though. But apart from that it’s very nice to use and quite nippy on my Pi 4.

I did read that the initial run takes a while but it took longer than I expected and I was beginning to think something had gone wrong when the window appeared. I was worried that it would be as slow every time, but on the second run, it flew. I took Chris Grandsen’s advice from another thread and turned off AdBlock and caching and it’s even better.

Apart from the ‘look and feel’ of it, can someone with privileged access to the beta browser run any comparisons? I noticed people were looking at start up times and loading www.autosport.com as examples of things to check.

I also tried Otter-VFP too, but it felt very similar to QupZilla-VFP but was slower. Is there any advantage to using the Otter-VFP over QupZilla-VFP?

How do they compare to the other browser that seems a bit elusive – OBrowse? Where did people get OBrowse? I’ve searched high and low and can’t find anywhere to download or buy it. I think I saw mentioned that Rcomp had something to do with it but when I checked their website, I couldn’t find mention of it anywhere. Their website does seem to be a bit out of date in places, so maybe they just haven’t added it yet?

Maybe a RISC OS browser comparison table, similar to the SVRSIG page on machines would be useful. Although I have no idea what you’d test to benchmark a browser on RISC OS. Would probably be more useful if it did something like ‘can it load x site and how fast’ with commonly used things like Google Mail, banks, Amazon and eBay and the like. So maybe a ‘can I buy things with it’ section too.

Maybe a RISC OS browser comparison table, similar to the SVRSIG page on machines would be useful.

Not sure about a browser comparison table. It’s mainly about whether a site renders and can be navigated.

Here’s a side by side comparison of Qupzilla and an early version of Iris (October 2019) before ROD took it on.
Qupzilla and Otter browser both use the QtWebkit (WebKit 1) library which is based on a fork of WebKit from 2013. Whereas Iris uses the latest gtkWebKit (WebKit 2) library.

Obrowser was an initiative that we (RISC OS Developments) started a few years back to allow people to make smaller contributions to fund things. It was primarily offered at shows on the ROD stand, although we did have a few CDs available separate afterwards. These could be obtained by contacting Richard @ Orpheus or Andrew @ R-Comp on the usual telephone numbers (the two main directors of RISC OS Developments).

Important note – ROD was never set up to deal direct with end users in terms of sales, because we don’t have the admin team / accounts setup / time to support all that. That is why ROD is funded by shareholders/investors/donators and gives its projects away for free (RISC OS itself, RISC OS Direct, Pinboard 2, LanMan98, Impression, TCP/IP and so on). The company tries to avoid wasting any funds on administration (except where absolutely necessary) in order to maximise what we can achieve.

OBrowser was/is a front end for Otter Browser written by myself. It uses Chris G’s Otter port (with permission) in a sub-folder called “engine”. The idea was that this folder could be updated with newer versions of Otter as necessary. The front end implements all the common RISC OS protocols (copy/paste, URL open wimp_message, ANT and Acorn URI protocols) to allow it to integrate with the RISC OS desktop and other applications (eg. email clients). You could also drag’n’drop URLs too it too, from text files. The idea was to create a front end that made Otter integrate better into the RISC OS desktop. It also did some pretty things whilst launching Otter, to mask the load time. ChrisG has since done something similar in the recent Otter/Qupzilla releases.

OBrowser purchasers will be the next, wider, alpha/beta test group for Iris, by way of thanks for their support.

Much of this is goverened by my available time. The community cannot help with this, I’m afraid, except by supporting R-Comp financially (ie. buying software/hardware) so that I can continue to dedicate roughly a third of my time (unpaid) to ROD work. I am rather a rarity in that I rely on RISC OS products/projects to feed my family and keep a roof over our heads!

Once the next beta phase of Iris releases, we’ll probably look again at OBrowser and how to re-shape it in light of Iris beta.

Right now, I’d ask for patience, and if you feel unable to support ROD financially (which I quite understand), just please let us quietly work on delivering the best (free) browser we can in due course.

PS, ChrisG – incidentally, that’s actually a ROD build of Iris as it has the graphics theme I created (from cc0 open artwork) for it. If you’d like to drop me an email, I’d be very happy to furnish you with the latest download – no point being a year out of date for no reason! :)

Not surprising that NetSurf does a better job with many modern web sites than modern WebKit, Opera, and Mozilla based web browsers do with the same sites

<spits tea across the room>

What.

What.

I’m sure it’s a typo. After all, this forum isn’t that complicated and last I knew Netsurf still didn’t handle the footnotes correctly.

This topic isn’t about Netsurf.

One year later. I can’t understand why Iris needs such much time to fix all the bugs. It’s a year and 18 months but still nothing. Or did I just forget something?

It’s a year and 18 months but still nothing.

I’m only replying to the suspected troll in case others are thinking similar things. 😉

I think it’s less about fixing bugs and more about getting something of that nature to work on an entirely alien system.
Do we have threads (in the POSIX sense)? Do we have X-lib or GTK or… ?

On Windows, WebKit uses Cygwin to bring Unix-like support. On RISC OS, it requires a little more work.

Plus, since a browser is a moving target, it’s going to be necessary to make an environment that can run WebKit, rather than just hacking it to work, because the more you mess with the original code, the less likely you’re going to be able to ever upgrade it to a newer release.

So, no, it’s really not just a case of “get this running under RISC OS and patch a few bugs”.

As for the time spent, how many people are working on it? The latest sources (for Big Sur) are 133MB compressed. That’s a lot of code, and that’s just the WebKit engine.

(I’m only replying to the suspected troll in case others are thinking similar things. 😉

I think it’s less about fixing bugs and more about getting something of that nature to work on an entirely alien system.
Do we have threads (in the POSIX sense)? Do we have X-lib or GTK or… ?

On Windows, WebKit uses Cygwin to bring Unix-like support. On RISC OS, it requires a little more work.

Plus, since a browser is a moving target, it’s going to be necessary to make an environment that can run WebKit, rather than just hacking it to work, because the more you mess with the original code, the less likely you’re going to be able to ever upgrade it to a newer release.

So, no, it’s really not just a case of “get this running under RISC OS and patch a few bugs”.

As for the time spent, how many people are working on it? The latest sources (for Big Sur) are 133MB compressed. That’s a lot of code, and that’s just the WebKit engine.)

I may understand it, but there are a million opportunities to do it faster. They may hire a QA staff for testing for example here https://u-tor.com/services/qa-staffing or more employers. It doesn’t make sense to time-stretch this process…

But I think I don’t know all the details, so I’m not very competent :)

They may hire a QA staff for testing or more employers.

Uhhhhhh……. 🤯

@Loraine When was the last time you paid money for a browser?