Quote:
Originally Posted by prusaks
review,
you are the best!!!
There are two functions that I would like to see for pdf files. First is as mentioned earlier x% zoom option. The second is regulating the darkness of font, some books have several fonts and the lighther ones are hard to read. It would be great if you could get it working on pdfviewer.
djvu viewer - your added feature has disabled everything else, can't do anything but go to next page. It has lost all the touch screen capabilities. The new feature is nice but can you make it optional? One book I opened had pictures and they were ruined to say the least with this djvu viewer. The thing that is missing in default djvu viewer is touch screen movement from right to left and l-to-r like in pdfviewer. Right now it allows to go up and down only.
wget - can you give me more information where it should be placed and so on? It does not have the .app extension and I couldn't find any file with that name on my PB302.
I could keep on writing about the things I don't like but for beginning this would be enough I think. By the way are these updates that you are making for fw 15.2?
Thanks mate.
|
Hi prusaks, thanks for your feedback.
pdfviewer:
I haven't had the time to commence the work on the pdfviewer but I will keep the comments in mind when I start going through the code. It is always good to have in advance a very clear idea of what one wants to implement and how. So the feature request and discussion should be at the beginning. So I will keep you updated as the work on pdfviewer commences and how it progresses.
djvu viewer:
But first let me comment on "your added feature has disabled everything else":
You're probably right with that.

but there are a few reasons to that which I will expand on briefly.
First: this was not a kind of final release for an updated and polished new version. Porting my modifications to the 302 models was done in a rush, so I noticed myself that some things were not working 100% in the updated version. But I'm still working on it and feedback is very important to get to a stage were the software is nearly perfect for all users.
The second problem is that my modifications were not based on the djviewer sources of the djvu viewer you have been using so far. But on the only one which has been published so far:
http://pocketbook-free.sourceforge.n...urcecode.shtml which seems to be rather old. So all additions I made were on that old source code, which may have removed some features you have begun to love in the original version. On my 903 I haven't noticed a great deal of missing features.
The biggest bug I've noticed in the ported 302 version is a problem with the zoom level from the quick menu. You can't select the zoom level with the touchscreen and none of the keys appear to have the function to go to left and right. So that is the first priority. This bug should be removed fairly soon.
So you say there was touch screen based panning in horizontal direction available in the original firmware version? Well, in the source code which I have used as a basis for my modifications there was only the vertical panning I believe. I can certainly add the horizontal panning as well, shouldn't be a big of problem. I will have a look at the source code.
Now, you said something about picture quality or that all pictures are gone. Now, this one isn't a bug. It has something to do with the way how djvu files keep the filesize low. I've mentioned it here and there a few times, so excuse me that I will keep it brief here now: Djvu files store the content of one page in different layers. The background and foreground layers and there is a mask to tell what is background and what is foreground. Now, pictures are usually stored completely in the background and saved with a high compression and low resolution (to keep the filesize very low), whereas text is stored with a less efficient compression so that the readablity of the text is excellent. However, since the text is usually only 5% coverage of the page you will still have only a small file size for the foreground layer.
Now, if you have a bad scan (I've included an extreme example of a bad scan) it will become very hard to read and enjoy it on the pocketbook. Because you won't have the full contrast. I've downloaded a few books from archive.org and they were literally unusable. The foreground was a dark gray and the background of the page was a light gray. This gave a very low contrast between the foreground text and the background of the page.
So my "improvement" was to be able to select which layer to be displayed. The default is now to display only the mask between the foreground and the background. Which will give the highest possible contrast ever. Also since the background is not rendered at all this reduces the processing power and might even lead to fast page flipping times and accordingly longer battery life times. However, since pictures are stored in the background layer, you will not see any pictures at all. So you may ask: What is the point then of calling that an improvement, it is not an improvement it is a bug. Well, the definition of a bug is as follows: A bug is a feature which cannot be disabled. So you would be 100% right if my new feature wouldn't be able to be deactivated. And I agree. So I've added "sensitive corners". Not exactly the corner but the first 25% of the screen in horizontal and vertical direction. So yes, fairly large sensing area. The functions are as follows:
double tap in the top left corner will change the render mode to the native way of rendering all layers: Foreground and background. So in this mode you will see pictures and might have a slightly lower contrast.
double tap in the bottom left corner will change the rendering mode to foreground, so you will not see any pictures
double tap in the bottom right corner will change the rendering to black and white. Also without pictures. This mode is the mode with the highest contrast but with no pictures.
Changing the mode is, as you can see, only a double tap away.
So you read in black and white mode to have the highest contrast and if you think you miss some important informations like a picture or so, you double tap in the top left corner. The page will be redrawn with full grayscale you can see the picture and then you decide if you want to continue reading with this slower rendering mode or if you want to switch back to the black and white rendering mode (double tap bottom right corner).
So the missing pictures is not a bug of my version, it is a feature of the djvu file format and pictures can be displayed in less than 1-2 seconds if you wish to do so while reading.
wget:
if you've never heard of wget before you most certainly don't need it now.

Let me explain: wget is a central function to allow any unix shell script to access webpages and get one webpage with all images etc. But not to render this page on the screen but to get the files. I've used it extensively in my newspaper application I've wrote for the german community. My newspaper application is not really an RSS Reader (as it wouldn't work for a random url) but it is adapted for each webpage. So the german community can get with in a few seconds a freshly rendered version of the news but only from a selected number of webpages. This will also work on the 302 and new recipes can be easily written for any given webpage.
My work is only in the userspace. So I'm not working on the firmware and I don't have the source code of the latest versions available. So this means that my improvements should be (within reason) independent of the firmware but I wouldn't be able to add anything which would require special permission. So anything which is not directly allowed by the system wouldn't be able to be added in userspace.
Hope this extensive and detailed answer helps.