h2g2 The Hitchhiker's Guide to the Galaxy: Earth Edition

I'd be happy to help test it.

This looks like a fantastic project!

RhoMuNuQ

You're on, sunshine!

The more, the merrier!

FM

"Sunshine"?

Thanks

By the way, is there any version of GuideDog that can be downloaded just now, or is it still being written?

RhoMuNuQ

It hasn't really started to be written. There's a proof-of-concept prototype written in VB6, which works (after a fashion). There's the functional spec/user manual which shows how it will work and what it will do.

When, and how, it gets written depends very much on the new version of Internet Explorer which is slated to arrive, in beta form, at the end of January. If it delivers what it is rumoured to, then we will have a very much easier job on our hands. If not, expect much sweat, tears and profanity between now and release.

OK ... I'd seen that manual.

I'm afraid I haven't heard much about the new Internet Explorer version, but I'll doubtlessly be using it when it arrives!

By the way, should I be listed under Testers?

RhoMuNuQ

You are now!

Thanks

Sweat tears and profanity? I hope not... Mozilla already does everything that we need to do so if the Internet Explorer thing doesn't work out we can always just switch our focus to writing for the Gecko engine. Even if the IE thing does work out we may find that it's easier to write for Mozilla. I found an interesting project on MozDev - a Visual XUL editor - but it's nowhere near ready for our purposes.

I need some condensed background on programming with Mozilla, if you're in a position to supply it. The IE thing may well fall through: it's too early to tell. However, as I have stated in other threads, there are very good reasons for preferring IE for now.

We have indeed had this discussion before. My intent was certainly not to rehash the previous conversation I just winced when I read "sweat, tears and profanity", that's all.

As a quick rundown, the Mozilla browser that you see when you start the application is just a set of XUL rules (Mozilla's user-interface definition format) and API calles wrapped around Gecko, the rendering engine. The interface that you can see is also rendered by Gecko, meaning that all you have to do to redefine that interface is to create alternate XUL files.

That's where GuideDog comes in. Given that the user interface can be redefined to have whatever widgets, rendering areas, etc. we want on it, it's (theoretically) not that hard to put the general purpose Gecko engine to use rendering GuideML pages. XSLT support is built into Gecko which makes it easy to reproduce the exact HTML code that the GuideML will generate once uploaded - all we need is an appropriate set of XSLT files for each skin.

Most of the application could be written in JavaScript, CSS and XML which is a good deal easier than writing in C or C++. If a new H2G2 skin were released or an old one updated, it would just be a matter of updating the XSLT files and downloading any new images.

If you want more info you can look at: http://books.mozdev.org/ which includes an online copy of "Creating Applications with Mozilla" from O'Reilly.

The prospect of a mozilla-based browser is very appealing, I admit. Unfortunately, there are the issues to do with distribution of the platform and its maturity, also the programmability of the browser itself.

I suggest we put out the stage 1 release using VB6 and IE6: this will do a subset of features (and it's almost written, really). After that, we take a decision regarding whether we go with Mozilla or IE. By this time IE 7 should be out, and we should be able to program the beastie using C# (which is MUCH easier than C++).

For now, Ion, please do carry on investigating Mozilla. Speak to MaW who is also doing the same.

Regards
FM

Ack! Distribution of the platform? C# would require the .NET framework which is 18MB, would it not? I'm not sure maturity is really a valid argument either, since Mozilla is based on the Netscape code base which has been around for longer than IE's.

I think what you've put forward would be the best course of action for now.

I shall get in contact with MaW.

"C# would require the .NET framework which is 18MB, would it not?"

If we use IE7 because it's written using .NET, then you can bet your life that the framework will be delivered with it.

"I'm not sure maturity is really a valid argument either, since Mozilla is based on the Netscape code base which has been around for longer than IE's."

But the editing platform API hasn't, and IE is based on Mosaic anyhow.

But anyway, I'm glad you think that the plan is OK.

MaW will be very helpful: he's determined to get it working with Mozilla so a bit of healthy competition will help.

I'm hoping (praying?) that IE 7 will be IE.NET, otherwise I probably will junk IE completely and go with Mozilla.

FM

I'm concerned that tying ourselves to the .NET platform will severely restrict the potential audience for GuideDog. I don't personally know anyone who has voluntarily installed the .NET platform and people will typically NOT perform system installs just to use your software. Plus a failed IE install can bugger your whole system, which makes many people hesitant to upgrade. If Mozilla fails installation you can just delete it and try again if you desire.

Plus for those with Mozilla (or any Gecko browser that's XUL compliant) already installed, GuideDog will be not much bigger than an image, packaged as an XPI file. Click to install, restart the browser and BOOM, GuideDog.

I concur with MaW. I'm determined to get a Mozilla version working. Competition is good.

I agree with everything you say. However, IE7 is rumoured to be built upon .NET so installing the new browser will install the runtime. Otherwise it's Mozilla.

You have the functional spec. Now go forth and Code!

Aha!

A perfect summary of what I've been trying to articulate since I became aware of the project:

This is a quote from a slashdot posting by JavaHacker.

"Installing Java is not the issue. [Microsoft] provided tools to their customers that rely on their VM being present for their corporate web pages to function. I know this because the company I am currently doing contract work for has a web application they purchased for content management of their shared project/business documents. It is written using Microsoft tools, and won't work if you have the Sun VM activated for your browser.

Their argument is valid, that this will cause problems for their corporate clients. It will cause problems whenever it comes out, because some of their corporate clients (or their customers) will not be able to view their web pages properly.

Delaying this rollout is not really going to help much, because most web application get updated when the application changes, not when the client changes. Their corporate customers are going to be very angry with them about this kind of problem.

I don't feel sorry for Microsoft, because they got themselves into this mess by trying to spin Java out of Suns [sic] control, and make it into a Microsoft specific version. Now they have been told to live up to their contract with Sun, and must pay the price for their behavior. I do feel sorry for their corporate customers who bought into systems designed around the Microsoft VM, because they were dumb, not culpable. They will end up paying part of the price for Microsoft's past errors.

Most corporate clients will have control of their desktops, and can make their internal users use the Microsoft VM until they can fix things. They can't make joe user on the internet do that, which is where things will break down."

I really doubt you're going to convince people to download IE7 (Beta) just so they can play with GuideDog, especially considering the 18MB download for the .NET framework plus whatever code is needed for IE, that's all.

I am trying my best to go forth and Code but unfortunately I have to put school and paying projects before GuideDog so I don't have an awful lot of time.

But what if Mozilla's editor simply *won't do* what we want it to do? Where does that leave us? Between a rock of installing IE 7 + runtime and the hard place of Mozilla's limited functionality, that's where

The only recourse we have then is to try to find a friendly Visual C++ programmer who can help us out. And there aren't too many of those.

If that were a problem, Mozilla wouldn't work at all. The user interface is rendered by the engine. The user interface we want is very similar to the default interface.

Well, anyhow, I get the feeling that MaW would very much prefer to strike out on his own. If you can follow up the Mozilla angle for this project then that would be good.

I don't really care what technologies we use for the editor providing the end product meets the requirements. Some of these requirements have not been spelled out, such as the ones regarding download times and availability of preloaded software. Perhpas this shoúld be our next step?

I've been tinkering a bit more and I can't see an easy way to make Mozilla display editable WYSIWYG text. I can think of workarounds but I'm not sure that's what we're after.

I have confirmed that XSLT support is built into Mozilla and is sufficiently stable for our purposes. XUL has also been standardised as of Mozilla 1.0 and there will be no more modifications (other than bug fixes) to the core functionality, which is what we would be relying on.

I'll keep digging.