Mini-review Sunrise 0.40-pre7

[hacker]

I just gave this version a dry run on one of my testing machines here, and it still suffers from the same problems that all of the previous pre-releases of Sunrise have suffered from... Java out of memory errors about 5 minutes into a fairly lengthy fetch.

It also is missing a lot of things that even early JPluck releases had as part of the base functionality. Granted, these are betas, but the basic functionality that is covered by its predecessor, JPluck, and the items supported by the Plucker document format, should exist.

Many of these useful options that exist in JPluck and were removed from Sunrise are in the FAQ, but it doesn't make sense to cripple the Plucker output for no apparent reason. Here's a few I just found:

1. All valid links are colored blue by default, invalid links are red by default. You can't customize these colors, or remove them entirely (by setting them both to black, for example)
2. No way to run Sunrise with optimized Java settings. The settings it uses by default are very braindead, and will consistently run out of memory on large documents with 1,000 or more links. This is 100% reproducable, and has been reported several times before.
3. No way to specify the Title or Author of the document
4. Can't set the Referer
5. Can't add UserAgent values to be used for the current or future document
6. Can't set Beam, Launchable, Backup bits on the output document (without having to add those with third-party tools at a later time)
7. No alternate images (major crippling functionality, especially for Plucker-base photo albums)
8. No way to increase the verbosity of the fetch session in orders of magnitude (1, 2, 3, and so on)
9. Logging is global, not per-document
10. No support for tables or setting bookmarks at fetch time
11. Only runs on Windows (JPluck ran on any machine that could run Java)

And now the bugs:

1. Sunrise crashes quite often when trying to add exclusions, wildcards, and regex limits to any document.
2. The fetch progress doesn't autoscroll the window. For large fetches, this is important, so you don't have to keep clicking the scrollbar to keep up with the fetch (yes, this is more of a feature-request than a bug)
3. Runs out of memory VERY fast on even moderately-large documents. Without the ability to create multi-megabyte files, it limits the functionality to "feeds" and short news sites.

Overall, its looking good, but it can't even remotely compare to the features, speed, capabilities, and functionality of his previous project, JPluck.

Take this with a grain, because I'm super-critical of most things, but this isn't the bees-knees or a replacement for JPluck. Laurens does great work, and his updates come out fast and furious.. I just wish that the existing JPluck users had something they could use that was better. This definately isn't it.

[Laurens]

Sunrise Desktop is not suitable (and not meant) for converting large documents, as we've discussed before. For this you need SunriseCL, which allows you to customize the VM memory settings any way you want. The root of the problem remains the core JPluck library, which makes wasteful use of memory.

The endgame is not to create a new Plucker tool, hence I chose to omit the Plucker-specific options. For instance, my own viewer can rescale images on-the-fly, hence no need for multiple versions of the same image. Also, the different link colors is something that could be implemented in Plucker itself, rather than hardcoded in the text record the way it is now. (Before drawing a link, check whether the RecordID is out of range and change the appearance accordingly. Should be simple to implement.)

That autoscroll feature is a good suggestion. Will implement it in a future version.

[bsquare]

To me, Sunrise is the closest thing I have seen yet to replacing AvantGo. The other options like iSilo, just aren't suitable for fetching news and RSS feeds. iSilo and JPluck are what I will use to create large documents, such as HTML Formatted books into iSilo, but that was never the intent of Sunrise.
Heck, the name "Sunrise" alone pretty much makes me think of the morning paper and getting RSS and News, not converting the O'Riley CD Bookshelf into Plucker.

With the latest release, the "always" schedule option allows the documents to be updated in realtime at hot sync. Correct me if I'm wrong, but the iSilo and others don't do this at all. I would like the ability to just sync at HotSync, and Laurens has answered this request with a definate "maybe"

I will agree with you on the "It'd be nice to have a Mac/Linux client" but doesn't that apply to a lot of apps with conduits out there still? Besides, Laurens has said that PPC will be the focus in the future, and how is the PPC sync support on Mac OSX? Limited at best. But, if this rumored low cost mac ever comes into fruition, I'll be hoping for better OSX support.

That said, I'm curious to see what direction Laurens will take this once he starts the viewer on PPC. I'd almost consider getting a PPC at that point, this may be the tractor app to pull me across the fold... maybe.

[Bob Russell]

Hacker has a lot more technical know-how on all these things that I do, so I'm sure he's right about the details. But I really like Sunrise also, and I've already completely replaced AvantGo with iSilox/iSilo and Sunrise/Vade Mecum.

And I am also tantalized by the unknown reader that Lauren is working on to fulfill that mysterious business plan! I'm not sure where he's going with it, but he promised a nice new twist, and I really like his work. It may not be a solution for everything, but I bet it's going to be well worth whatever he decides to charge for it!

But what drives me crazy about them all is that you can't automatically open document pages at n-lines or n-screens paged down. I hate paging through all that left hand column comment or header info to get to the "real" content on a page. RSS + printer friendly link replacement helps sometimes, but I really just want to be able to open pages and linked pages of a document by default at the location I want to read from. But, hey, you can't have everything!

[hacker]

Originally Posted by BobR

But what drives me crazy about them all is that you can't automatically open document pages at n-lines or n-screens paged down. I hate paging through all that left hand column comment or header info to get to the "real" content on a page. RSS + printer friendly link replacement helps sometimes, but I really just want to be able to open pages and linked pages of a document by default at the location I want to read from. But, hey, you can't have everything!

Sure you can!

Sitescooper has been doing exactly what you're looking for for several years now. Give it a try!

Sitescooper automatically retrieves the stories from several news websites, trims off extraneous HTML, and converts them into formats you can read on your Palm computing device for later reading on-the-move. It maintains a cache, and will avoid stories you've already read. It can handle 1-page sites, 1-page with diffing, 2-level and 3-level sites, and it's very easy to add a new site to its list.

Even if you don't have a Palm handheld, it's still handy for simple website-to-text conversion, and offline HTML reading. For example, here's some screenshots of an iPaq displaying sitescooper output.

[Bob Russell]

Originally Posted by hacker

Sitescooper has been doing exactly what you're looking for for several years now. Give it a try!

Thanks Hacker. Yes, it looks wonderful in terms of what you can do to choose what you get out of it. But when I tried it before, I gave up because support seems non-existant and I got version incompatibilities between sitescooper code and iSiloCX. See https://www.mobileread.com/forums/sho...owpost&p=11412

Kind of figured it wasn't the sort of thing I wanted to mess with if I was going to have to dig in that much. (I like the easy solution for this sort of thing because I spend way, way, way too much time on it already!)