Question to the experts: What is better for the “html-ui” framework: components on divas or on html standards?
There is an idea: Port for (not yet released) gwt 2.0: flex4 components (along with the maximum possible fxp1.0). Well, let's say in the com.google.gwt.user.client.ui.flex4 package.
First, let's say, I really like fxg itself, and indeed flex4 itself, and besides, the existing gwt-ui does not give me complete freedom in fast implementation, at least 80% of my projects. Well, all existing libraries weigh too much, and they slow down no less. And I don’t like their semantics at all.
It is required to port 100% cross-browser, essno (well, as the flash-plugin itself renders).
')
I make up pretty well (well, okay, the rest), but I have never been involved in benchmarking and generally technical aspects of layout rendering, and there are differences in my team, so there are a few questions for the community:
1. What is better to use in the custom component, using, say, checkbox or radiobox logic - divas with appropriate logic or standard html components, applying css2- "3" to them and framing them with divs, if necessary (well, for further beauty, which may be missing in css)? It is clear that [input type = "textbox"] / [textarea] are necessary in principle for the logic of edit components (although many browsers seem to support editable, but I have never worked with this property), but you can make a button, etc. in several ways, but it is not clear which is better, or is it really pofig?
2. What is more appropriate to use, perhaps a large css piece, or a small js supplement? Ie, in gwt2, we copied layouts, because they reacted inadequately to onResize, in the end they chose pure css; Where is the golden mean for most customers? In general, how best to choose the balance: css (2- "3") - js / cross-browser (in terms of minimalism of the written code, because gwt loads different for different useragent, if necessary), which beats the performance of technology more? Ie, if I speak through js to change the width, the render changes (on average by browsers), and then the css-property, or does js change the style first, and then does it render? Or, in general, everything is different and you should not be bothered by this, or are there any aspects?
3. How widely should html semantics be used? Ie, in fact, you can limit yourself to div tags only, but you can also use span, li and so on in “necessary” semantics instead of “div with style” (and even take it from html5, and if so, to what extent)? Ie, how will it be more convenient and more correct for the final developer of the application on my components? And how will it be better for the web (clients + search engines)? And how can this affect the speed of rendering (if div + style, instead of standard), well, let's say with 1000 such components on the page?
Well, your wishes and suggestions will not be superfluous.
First, let's say, I really like fxg itself, and indeed flex4 itself, and besides, the existing gwt-ui does not give me complete freedom in fast implementation, at least 80% of my projects. Well, all existing libraries weigh too much, and they slow down no less. And I don’t like their semantics at all.
It is required to port 100% cross-browser, essno (well, as the flash-plugin itself renders).
')
I make up pretty well (well, okay, the rest), but I have never been involved in benchmarking and generally technical aspects of layout rendering, and there are differences in my team, so there are a few questions for the community:
1. What is better to use in the custom component, using, say, checkbox or radiobox logic - divas with appropriate logic or standard html components, applying css2- "3" to them and framing them with divs, if necessary (well, for further beauty, which may be missing in css)? It is clear that [input type = "textbox"] / [textarea] are necessary in principle for the logic of edit components (although many browsers seem to support editable, but I have never worked with this property), but you can make a button, etc. in several ways, but it is not clear which is better, or is it really pofig?
2. What is more appropriate to use, perhaps a large css piece, or a small js supplement? Ie, in gwt2, we copied layouts, because they reacted inadequately to onResize, in the end they chose pure css; Where is the golden mean for most customers? In general, how best to choose the balance: css (2- "3") - js / cross-browser (in terms of minimalism of the written code, because gwt loads different for different useragent, if necessary), which beats the performance of technology more? Ie, if I speak through js to change the width, the render changes (on average by browsers), and then the css-property, or does js change the style first, and then does it render? Or, in general, everything is different and you should not be bothered by this, or are there any aspects?
3. How widely should html semantics be used? Ie, in fact, you can limit yourself to div tags only, but you can also use span, li and so on in “necessary” semantics instead of “div with style” (and even take it from html5, and if so, to what extent)? Ie, how will it be more convenient and more correct for the final developer of the application on my components? And how will it be better for the web (clients + search engines)? And how can this affect the speed of rendering (if div + style, instead of standard), well, let's say with 1000 such components on the page?
Well, your wishes and suggestions will not be superfluous.
Source: https://habr.com/ru/post/74004/
All Articles