From: Michiel D. <mi...@ti...> - 2010-09-28 09:54:10
|
With my own OS project, phpList, I've just spend easily a few months with a coder and a CSS-er to revamp the dramatically outdated interface. I can tell you, it's not for the faint hearted, but once you get the basics down, it's really worth it. phpList has about the same age as mantis. I have no in-depth understanding of the mantis code and how it separates the business layer from the display layer, so maybe many of these things are already in place, but I would advice to approach it using widgets. Do not go through designs page by page. Instead you build a library of interface elements that each work out their own way to display things. Then in the page code, you instantiate a widget and add content to it and call a final method to display it's output. Also, with current web technology, tables are out. Everything sits in a div or span with a class. Name the classes by "function" and not by display aspects, eg class="action" and not class="leftmenu" (who knows, the menu might go to the right) Also, stylers love "UL"s for listings. Basically, the coder determines the elements that go in a page and the styler determines how they look and where they should go. I have for example a "listingObject" that in the code I add elements to with columns. At the end, I say "give me your output". It used to output the entire listing in a table, and now it simply returns divs with classes so that the styler can get his hands on it. Michiel On 28/09/10 07:58, Robert Munteanu wrote: > To follow up on my comments, I think that we - as a development team - > should come up with a clear roadmap related to the redesign of > MantisBT. > > While stating that it will be difficult to achieve is nothing but > correct, it's not bringing us closer to a resolution of what I > perceive to be the greatest weakness that we have. > > I must admit that I have not worked with the rendering APIs at all, > but I think that such a task would at least involve: > > 1. Redesign of the view issue page ( kudos to Oskar for picking up this one ). > 2. Redesign of the view issues page > 3. Redesign of the report issue page > > 4. Decision on how to implement changes . Should we refactor the core > APIs to support templating for these changes? Yes, it makes sense, but > on the other hand it might not be required. A template promotes good > coding practices, but we should weigh the benefits of making this > change and then implementing the design as opposed to the > disadvantages of yet again postponing a much-deserved redesign. > 5. Implement changes in the core APIs, if needed. > > 6. Implement changes on the view issue page. > 7. Community feedback > 8. Fix, and finish implementation. > > All that we - again , as a development team - must do is 4 - decision > on how to implement the changes , and perhaps assist in 5 - implement > changes in the core APIs. > > So my suggestion is > > 1. Evaluate what the best way of implementing these changes is. > 2. Create any needed support in the core APIs . > 3. Steadily apply changes to the design, as it unfolds. > > I don't necessarily see adopting a templating engine as a > prerequisite, simply because it seems a huge task in itself. > > Thoughts? > > Robert > > On Mon, Sep 27, 2010 at 10:23 PM, Robert Munteanu > <rob...@gm...> wrote: > >> On Mon, Sep 27, 2010 at 8:55 PM, John Reese <jr...@le...> wrote: >> >>> Paying MantisBT developers for their work is just messy IMO, esp with >>> international regulations and taxes involved, and only serves to >>> complicate the situation. I would actually greatly prefer that you >>> contribute your employee's time to learning the MantisBT codebase, >>> joining us in discussion of the features, and then developing the >>> features in the open and working towards getting them accepted into our >>> repository. Time is always the biggest constraint on our existing >>> development efforts, and contributing developer time and effort is IMO >>> the best way to get things done. >>> >> A huge +1 on that. One or more developers willing to work together >> with us to refresh MantisBT's look is the best we can hope for. >> >> >> -- >> Sent from my (old) computer >> >> > > > |