Gecko Runtime Environment Overview
Sunday December 22nd, 2002
David Hallowell writes: "Doug Turner has published a message to n.p.m.embedding about the Gecko Runtime Environment (GRE). The GRE will make life easier for people who want to embed Gecko as well as meaning apps like Phoenix, Mozilla and Thunderbird will be able to share common libraries, saving on disk space."
While the idea is interesting, is there really any good reason to do this on desktop operating systems? After all, disk space is hardly a constrained resource these days.
Doesn't the sharing of Mozilla libraries by different Gecko-based applications expose the user to a lot of problems with version mismatching between application and GRE?
There will be support for haivng different versions of GRE on the system, so that applications that need a more recent version of GRE can install the later version.
The main benefit of this is that applications can use gecko but be a smaller total download... e.g. they could offer a smaller download for those that already have the GRE installed, or if they're using a net installer then it can detect whether a GRE is installed and if so reduce the total download size.
This will also mean if AOL do move to a gecko based solution there'll be a large number of machines that already will have the GRE installed on them, and this may encourage more people to use gecko rather than IE or at least make it an option
#6 Size of the download: does it matter that much?
by PaulB <firstname.lastname@example.org>
Monday December 23rd, 2002 12:17 PM
If GRE is being developed solely for the purpose of reducing the size of the download I rally do not see the urgency. As the transition to Broadband Inernet connection proceeds the size of the download no longer matters. With the cable internet I have here I download a new nightly each day and rarely does it take longer than 1-1/2 to 2 minutes. I know many still depend on a connection <53.3kbs but isn't this a steadily shrinking group?
Anyway if GRE is being developed partly to reduce the size of Mozilla to encourage more users to download and use Netscape, this might work, but there are much better ways to accomplish this. Here are a few suggestions off the top of my head:
Rather than try to reduce the size of the download to as the main method to encourage users to download and use Netscape: 1. AOL when it ships out the next batch of cd's could clearly label them as including a free copy of Netscape 7.
2. Netscape might develope deals with computer makers such as Dell and Apple to have Netscape included as the default browser. Many users will only use the browser included with the computer. Some are even afraid of messing with the software as installed by the OEM for fwar of messing up some setting. If Netscape does not start to develope deals with computer manufaturers to install Netscape 7 as the default browser some people will never even bother to install it. Some literature could be developed to offer OEM's to include in product literature stating many of Netscape's features, especially its security features.
3. Netscape could evangelize its browser and mail features to corporations. Mail has reached a state of development (featurewise [namely its spam and other filters]) that reviewers are beginning to stand up and take notice. Some corporations are tired of the level of security in Microsoft's internet Web and Mail software and are looking for viable alternitives. Netscape need to inform that it can offer a viable alternitive to Microsoft's internet offerings.
Yes you heard correctly...I said Netscape and not Mozilla. Most end users do not want to use a developmental software (i.e. Mozilla) and even some corporations even have policies forbidding the instalation of beta or developmental software. Netscape has undergone a more thorough testing before it is released (OKAY, OKAY there was NS 6.x) and because of this people precieve it to likely be less buggy, less likely to cause data loss, and of suffient quality to install on a corporate network. This same perception does not hold for Mozilla. Mozilla is viewed as software still under development, not finished software like Netscape 7.
"As the transition to Broadband Inernet connection proceeds the size of the download no longer matters."
Even if all users had high-speed connections, download size would matter to the distributors who pay thousands of dollars for servers and bandwidth.
While disk space has increased at an incredible rate, memory size has increased at a slower rate. By using a single copy of the gecko libraries, each of the mozilla based apps you run can share the read-only data from those libraries (code, const strings, etc). If you only run mozilla, this probably doesn't seem like much of a win, but if you run multiple apps (eg. mozilla, phoenix, standalone moz mail, komodo, etc), it will make a difference. For the previously mentioned apps, you would see even more of a saving by using the XUL Runtime Environment (once it is ready).
#5 Re: When will Netscape move beyond 1.0.x
Monday December 23rd, 2002 10:10 AM
It's mainly to allow applications to build against a stable, known release of Gecko without requiring they go off and grab, build and bundle it for themselves. It also allows us to consider releasing an SDK of headers and libs which work against a particular GRE. It should be a very useful feature for embedders, especially those who don't want to subject their users to unecessary 5Mb downloads when they already have a GRE.
#3 Set GRE directory using envirenment variable
Sunday December 22nd, 2002 11:18 PM
Under unix at least it is common practice to have three levels of configurability: - Per system (/etc/grerc) - Per user (~/.grerc) - Per session (environment variable: GREHOME ?) You didn't mention the last one, but I suggest you add it, even if it only helps tremendously during testing.
Succes with the project!