Brian J. Murrell wrote: > On Mon, 2008-07-14 at 17:04 -0400, Eric Bishop wrote: >> It is ironic that the LuCI team decided to make an announcement >> regarding their project today. I have also been working on a new >> (open source) web interface for Kamikaze called Gargoyle, and am now >> releasing the first beta version, which can be found at >> gargoyle-router.com. > > Another. ~sigh~ > > I'm all for choice, but too much choice possibly means an unnecessary > division of labour and unfortunately all Open Source projects suffer > from a shortage of labour. > > I wonder how much better and more complete one (or two perhaps) web UIs > for OpenWRT would be if the resources were pooled into a common project > that kept all of the stakeholders happy. >
In a perfect world, yes, but pleasing everyone is hard. Existing projects are simply not suited to what I want. I looked carefully at existing projects before starting Gargoyle and concluded that to do what I wanted would require re-writing almost everything, so I started my own project. > >> Currently it is designed to run on top of Kamikaze 7.09 and not the >> trunk, but as soon as another stable version is released it will be >> engineered to run on top of that. > > So you will only remain compatible with released versions? How much lag > do you expect after a release becomes stable before you will have your > UI working on it? > You are right, there will be a lag -- probably about a month. However I fail to see how this is a major problem. It makes a lot more sense to develop for a stable release than for something that can (and does) change at a moment's notice. Any key must-have features/bug-fixes in the trunk can be added on top of the Gargoyle versions released for older distributions. This is how the new UCI is currently incorporated. >> Gargoyle takes a very different philosophical approach to interface >> design than X-Wrt or what I've seen of the new LuCI. Both X-Wrt and >> LuCI seem to be designed with the goal of providing the absolute >> maximum functionality possible. > > Which is a good thing. With the hackabilty of OpenWRT, people are doing > all kinds of neat things to it, a lot which less-power-users might like, > if there were UI to configure it. Right, but only if that UI is easy to understand. Just because a feature CAN be configured via a web interface does not mean that it is accessible to your average user. In order for an average user to be able to use a feature it must be relatively easy to 1) understand what it is/does and 2) understand how to use the controls avaialable to configure it. Even if the feature is available via a web interface there is a barrier to usage unless the feature is properly labeled and the controls are easy to use and to understand. >> However, this often comes at the expense of making the interface more >> difficult to use, and can turn off novice users. > > Indeed, there should be a simplified interface for the simple use-cases, > but also a more advanced interface for power users. But there is > nothing wrong with those both being available in the same UI. They are > not mutually exclusive. Again this is true in a perfect world. It is POSSIBLE to create a user interface that incorporates all desirable features for power-users and is still very user-friendly. However, as you note above, there is always a developer shortage, and there is ALWAYS another new feature that can be implemented. Therefore, there is always a tradeoff in how much time is spent developing each individual feature so that it will be user-friendly and how many features you have. Existing projects devote almost all of their time to implementing new features as opposed to spending the time (and it always does take a lot of extra time) to make features very user-friendly. The result is that existing projects have tons of really cool features, but many aren't accessible to the average user, and need drastic changes to make them accessible. Hence, Gargoyle. Eric _______________________________________________ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel