Ktn

I used the 1.0.x versions of Thunderbird during several months, I don’t really remember when I first installed Thunderbird, I was previously using the Mozilla suite and moved to Firefox and Thunderbird at the same time. Everything was almost smooth and good with the 1.0.x, until January this year when I updated from 1.0.7 to 1.5.

I said almost smooth above because I’ve suffered the consecuences of the Duplicate entries appear in feeds bug.

Thunderbird 1.5 continues to have a lot of good features as an RSS/ATOM aggregator, bear in mind that I’ll mention here only this support, because as an NNTP, and mail client it’s excellent.

The way you can group the feeds, the way it shows the name of the feeds and their icons, the list of entries and their icons, all are so descriptive and well done. I like a lot the field to search for the posts by title or sender.

The RSS subscriptions now default to load the web page for each feed entry. From Thunderbird 1.0.7 I already had the “Show the article summary instead of loading the web page” checked for all of the feeds I’m subscribed to. When I updated to 1.5, I was subscribed to 84 feeds, had to verify one by one that the mentioned option was checked, a waste of time indeed. The big problem is that for several feeds it doesn’t want to pick up the summaries again, hmm.

Seems that the XML parser is broken in some way, I’m pretty sure is broken for RSS feeds. Sometimes it just grabs one blog entry when other feed readers retrieve more, or maybe it just doesn’t goes and retrieves the feeds again; something happens internally that it stops checking the feed. In some other occasions it says that the feed is not valid, but according to Firefox and RssReader the XML is valid and parses without a problem. So, yes, if I were using only Thunderbird to keep track of news feeds I would have already lost a lot of blog posts.

Looks like it doesn’t pay attention now to the dc:date field. After I installed Thunderbird 1.0.7 it got 22 blog posts from one feed but the Date column shows 1/15/2006 11:23 PM for all of them!! This is pretty bad, obviously the time isn’t the same for all of them, this is a regression.

The Manage Subscriptions has some odd bugs. The RSS Subscriptions window shows me a folder without its feed and I can see the feed and its entries in the Thunderbird left pane, how am I supposed to modify the settings for that feed? For another folder it showed me 3 feeds with the same name and allowed me to edit the feed properties, so I decided to remove 2 of them and now I can’t edit the properties! In fact now it’s showing me 2 feeds with the same name in there but can’t edit any. Pretty weird and annoying.

One thing I should applaud is that at startup it uses several threads to retrieve the feeds and the updates are pretty quick, without disruption of CPU usage or memory consumption.

One problem that has been so mentioned here and there is that it has some memory leaks that I believe come from the Firefox codebase.

Right now Thunderbird is using 150 MB of memory, and this is due to the memory leak, becuase yesterday it was using around 70 MB consistently. And when selecting a feed it uses some amount of CPU for less than 2 seconds, with a peak of 70% of CPU, which is more than excellent. Now that I selected another feed the memory usage is 88 MB, seems it depends of the selected feed and how many entries it has. The biggest one, with more than 67,000 entries, causes Thunderbird to use 139 MB of memory.

The last thing I have to say about this application, that actually talks very good about it is that is the only one that is handling the biggest number of posts, the older post which is when I did the last backup is from July 15th, and the total number of posts may be around 126,000 !! Using 231 MB of hard disk space.

Something that I don’t like too much is the way it stores the feeds on hard disk, looks like when the name of the feed is big it creates a folder with the name aa8f7dc2.sbd that doesn’t have any contents in it! There are a couple of files with the same name as the directory but at the same level as the directory itself, one without extension and the other file with an extension of .msf. So, in the end there are 3 objects per feed, an empty directory and 2 files.

I’ll give Thunderbird a chance, I’m going to backup all of those files, in addition to my newsgroups and mail files, will uninstall it, remove any of its files, user files, and then I’m going to install it again, restore the mail and newsgroups stuff but add the feeds one by one being sure that for all of them it’s retrieving the summary or whole contents of the posts instead of loading the web pages.

Wish me luck!

I really appreciate the good work the developers did with this tool, I wish the issues I’m having are fixed in the near future.

Kudos to the developers and to the Open Source effort around the world!

This time I’m going to talk about RssReader which was the first news feed reader I started to use a couple of years ago.

RssReader is one of those that is also simple, but pretty good at doing his job. There are not much bad things I can say about it but I will mention them here.

It uses the .NET framework and obviously the IE engine. I have 93 feeds in here, and due to the use of the IE engine, sometimes I get IE popups opened, really annoying.

I’m using version 1.0.88.0 which is 2 years old, the latest one seems to be 1.0.91.0 that is a beta released for testing and I haven’t downloaded it because of it, maybe it’s better than the one I’m currently using but not sure if stable enough, perhaps I’ll give it a try in some more days.

I would like to see easily the number of totals posts versus the number of read and unread per feed and in summary for all of the feeds.

I really love the ease of use of the tool bar, the option to check the feeds, to add feeds, to add groups, to edit the properties of a feed, the option to clear the headline history, the pull down menu to select the type of the headlines I want to see in the listing pane for example, the new posts, the unread, the read ones. The date filter is also usefull and what to say about the keyword field to search the headlines, just awesome!

I like that it can be minimized to the system tray, this can be done with the others also but with RssReader is easier, maybe not, but there’s something special. I like a lot the way it shows the popup with the headlines of the entries after it finished checking a round of them.

Another thing so unique and that I appreciate a lot is that you can customize the size, style, color of the font for each kind of item in the reader, for example, for groups whether read, unread, etc.; for feeds read, unread or failed; for headlines read, unread, new, important, etc.; for the previews; that way you can easily identify what you want.

It has several minor issues that can be improved of course. One of them is that I don’t know how it does the scan of the feeds, seems like if it uses only one thread and takes a lot of time to go through all of them; when it’s started the icon in the system tray doesn’t respond for several minutes, and if one of the feeds is not reachable or there’s a problem with the feed it shows a little window (modal? I forgot how that’s called) telling that it failed to read the ATOM feed, it doesn’t tell which feed nor if it’s really an ATOM one, but it fails with the same message always. If I don’t click on it then it won’t continue scanning the rest of the feeds, sometimes it does it, it’s kind of strange.

I don’t know why it does not remember when I right click on the system tray icon and click on the “Sound on” to disable it, whenever I start it I can see that the option is selected again, hmm.

When talking about feeds and how it shows them I’m going to mention that it had troubles when creating the html to display the feed for the IBM internal blogs, the published date sometimes wasn’t in the place it should be, and the entries got unaligned in rare cases.

Something that I learned about it is that all of the contents of the feeds are stored in the storage.xml file and the HTML files stored under the HTML directory are created only when that feed was selected, so the page shown had to be recreated again and again; this applies also for the groups.

It currently has around 12,000 entries using 9.6 MB of hard disk space and obviously using other 9.9 MB for the HTMLs and icons for each feed. It doesn’t have much right now because I archived the old entries 2 or 3 weeks ago, I’ve had around 70,000 entries previously with the proportional hard disk usage.

I was forgetting to mention that when it has more entries it takes it more time at startup to parse the entire xml, and the memory usage climbs up to 300 MB. Something for which the culprit, I think, is the .NET framework.

In general after startup RssReader is responsive and quick, and doesn’t needs excesive amounts of memory. When selecting a channel or a group it eats more memory, and CPU also, but when it’s done everything returns to comfortable levels.

For sure I’ll keep RssReader, it does the job and provides me with several simple functions that I need. I must say that the developers did a great job. I really look forward to the next version that is not a beta for testing.

Keep up the good work !! I’m eagerly awaiting for the next release

Newzie is one of the most complete RSS feed readers out there, it has a lot of functionalities that I believe I won’t mention, basically because I don’t use them!

Indeed, it is one of those tools with a lot of things in it that it just got bloated. I will hardly use even the 10% of the functions it provides. I don’t know why but this sounds familiar, with the huge difference that Newzie is free.

I’m using the latest version for a long time, it’s the 0.9.91 Beta. I’m not sure if it uses the .NET framework or not, I couldn’t find anything in their web page, but for sure it uses the IE engine, something I don’t like.

I should say that Newzie is pretty good handling the feeds, even when some have parsing errors in other readers, Newzie parse them correctly. This is something good that really surprises me. I don’t know if the developers actually created an algorithm to detect the wrong things in the feed to fix them or if it was added without having the intention to.

The thing about giving each channel a value, like 1 cent worth, or a buck worth, 20 dollars worth, etc., is nice. And the colors in the listing are good also.

One of my issues with it is that most of the times it forgets when I selected the “Calendar Navigation” for a channel. Another issue is that by default it uses the entire screen and there’s no way to have it maximized but reduced in dimensions, hope you can understand what I mean, I like to always see several windows at a time and Newzie is kind of intrusive.

As I said, one of the things I hate the most is that it uses the IE browser engine, and you may think that’s the worst issue that can happen to someone but not ! There’s a problem with Newzie that is really bad…

If it doesn’t crashes during the first minute after I launched it, it stays consuming 99% of the CPU!!! Yes, pretty pretty bad, I think it doesn’t consumes the 100% just because it can’t.

I can be sure if it crashed in the very first minute if the machine is responsive, if it doesn’t crashes I have to open the Task Manager and end its process.

In the window I see when it crashes I can read the common text when an application crashes in Windows, which is:

Newzie - Information Aggregator has encountered a problem and needs to close. We are sorry for the inconvenience.
If you were in the middle of something, the information you were working on might be lost.

There’s a link at the end to see more information, and when clicking on it I can see the following:

Error signature
AppName: newzie.exe AppVer: 1.0.0.1 ModName: lpdll.dll
ModVer: 1.0.0.1 Offset: 0000a4e2

Another link appears in this window to see the technical information, but it’s useless for me; and I won’t provide that information to the developers either because some of the contents as hexadecimal values that are in there may be something from the feed of the internal blog site in IBM.

By the way, when clicking in the Close button in the window that notified me that Newzie crashed I can see an Error window containing the following:

Newzie.exe - Application Error
The instruction at “0x7c2821b2″ referenced memory at “0x0000001c”. The memory could not be “read”.
Click on OK to terminate the program

So, I’m stuck. It never works correctly. Surprisingly it uses just a few MB of RAM in comparison to the readers that explicitly mention that the .NET framework is required.

It is using 168 MB of hard disk space for the nearly 24,000 entries it has right now, from 44 channels. And looks like it maintains indexes because, apart that I see an “Indeces” directory, there’s a lot of hard disk read and write when Newzie is starting. Yeah, I know, I probably can configure it to not go and check all of the feeds at startup, but I really need it to check everything at the moment.

I’ll try the “Backup Current Environment” feature and remove it until there’s a new version released. Then I’ll see if the crash or the high CPU usage are gone.

This RSS/ATOM agrregator has several good things but also a couple of very bad things that I’ll try to mention in this post.

I feel that I have a lot of items to point out about GreatNews that I’m fearing this post will be quite large, hope it isn’t.

In GreatNews I have only 4 channels, with a total of… hard to know in this case, but my guess is that less than 5,000 entries that use 8.43 MB of disk space (according to the newsfeed.db file).

The interface is good, you can add channels easily and group them. In this case I’m able to configure it to show me just one post when selecting a channel in the left pane. Yep, you guessed it, I had to do that because GreatNews was also getting unresponsive for a few seconds when going to a certain feed, it was configured to show 10 items at a time. At first I thought it was using the .NET framework but their download page says it doesn’t need it. It’s strange because when right clicking on links the context menu is the one from IE, and when seeing the properties the window is almost the same as the one in IE. I said almost because I don’t remember them very well, I hardly use it.

Something that is good but that almost all of the readers provide is that when a feed is broken or can’t be reached the name of the channed gets red, so you can know easily which channels are not working at a specific time, and shows you a tooltip about the error, but isn’t helpful at all.

I never tried the “Label This” and “+Del.icio.us” options since I don’t need them so I can’t comment on them. It provides several styles for the reading pane, which us good, but there is not much difference between them.

Something that almost convinces me about it not using .NET is that when idle it uses only 14 MB of RAM, which may be a point in favor of GreatNews. But to be fair this is the RSS reader with the less workload of all.

I would like to see an easy way to discover the number of total, read and unread posts of all the channels, you can see the properties for each channel one by one, but not for all of them at once.

The problem with GreatNews is that, for the feed of the internal blog site in IBM, it shows a lot of duplicate entries and it also shows entries with the wrong title (taken from another post), the wrong author, and it even tries to predict the future!! Yes, indeed, I see a lot of posts supposedly created on April 30th, 2006. As you can imagine, when I’m trying to see the latest posts in the listing pane I’m seeing just garbage posts, or in other words, posts that were in fact created but that shouldn’t be there. Well, also shows duplicated and “mixed” posts.

It’s a shame that this tool can’t handle the feeds correctly, I won’t be able to try it full force until that is fixed. Have to say that the feed where GreatNews is failing so bad is (was for the already reviewed readers) already added to the other feed readers, and this tool is the only one that shows this problems.

The last bad point to mention is that when starting the tool it doesn’t remembers my preference to show the listing pane by default. I have to go to “View” -> “News list” over and over again.

I’m using the Build 1.0.0.364 of the 1.0 Beta. Time to uninstall it.

Now is the turn of SharpReader. I’ll mention up front that I’m using what appears to be the latest version (0.9.6.0) as of today and that it wasn’t put into the pressures as the ones in which RSS Bandit was.

SharpReader is simple, I’m not sure if it was created with a minimalist goal in mind or not since I’m not a fan of minimalism and actually I don’t know what defines minimalism.

I like the option to see the “Feed Properties” at the bottom of the left pane, I have this little section active because I’m only subscribed to 5 feeds, 2 of them are for testing purposes and another one just have 5 posts, it’s holding only 5,337 entries that use 10.9 MB of disk space.

It would be good if SharpReader can provide a little button or icon at the bottom of the left pane to be able to toggle on or off the “Feed Properties” section, to avoid having to go to “Tools” -> “Feed Properties”, I’m mentioning this because in the case I’m subscribed to the 102 feeds as in RSS Bandit I may have the need to having a look to most feeds as possible in the left pane but quickly take a look to the properties of a given feed.

Another thing that I liked but doesn’t work so well is the fact that by default it shows you as additional entries under a given post the thread of entries for it and sometimes the links to references to web pages or images contained in the post, somehow handy, but needs to be polished.

A bad point, there’s no way to create categories or folders of feeds. Yep, I cannot group the feeds so that I can go to the ones I’m more interested at the moment, this kept me from adding the whole list of subscriptions into it.

In addition, there should be an horizontal scroll bar to be able to see the entire name of the feed and unread vs. total number of posts for that feed.

I’m still unsure how the “Subscribe” button should be used, it was always disabled. I have to go to “File” -> “Open RSS Feed” to start the process of get subscribed to a feed. And yes, I know that the Ctrl+O keyboard shortcut is available, but since I only added 5 feeds I never got used to it.

This is fun! I forgot that the Search field instead of searching in the entries that are locally stored goes to feedster and retrieves those posts containing the searched word from several feeds, and then the “Subscribe” button is enabled! Unfortunately is not obvious if I’m going to get subscribed to the feedster search string or to the feed of the selected entry. In any case my intention is to search in the posts of the feeds already in SharpReader, I may need to look into something I previously read.

A hidden issue is that it also uses the .NET framework, but was not critical because as I already said I have just 5 feeds in here and the performance issue isn’t seen because when selecting a feed in the left pane, no entry is shown in the reading pane. Something that starts to be of concern is that it is using 72 MB of RAM while idle.

A good thing is that when it’s starting it doesn’t consumes so much CPU as RSS Bandit (I think I forgot to mention this in its review).

I think that’s it, I didn’t used it too much so I’m unable to say more about it.

In summary SharpReader is pretty good in doing the job, but for guys like me with a lot of big and real needs, which I think I’ll have to point out in the summary of reviews, it still needs some little changes and a selection of a different framework.

The first RSS/ATOM aggregator I’ll talk about is RSS Bandit which looks really good but has a big problem, it uses the .NET framework.

It has for sure a lot of good points, but you’ll see that my requirements are perhaps so extreme or beyond the edge that this application isn’t good for me, or better said, for my machine.

The look and feel are just excellent, I must say; the way the buttons to add a new feed are placed and designed is very good, the easy it is to add a new folder/category to place the new feed in there is awesome, the icons are good, the “Feed Errors” special feed is something that provides an easy way for geeky people to go and look why some feeds are failing.

The reading pane has a good design by default. The pane that shows the list of entries is quite good, but the default colors didn’t worked for me, I had to modify something in there, I think I reduced the font size and modified the color of the entries not yet read to blue, something like that, I may not recall what I really changed. The missing thing there is that it does not have the functionality to differentiate new unread posts (those retrieved in the last check of the feed) from the old unread posts, that’s something that Thunderbird does pretty good.

The bad points
Almost all of the bad things about the tool are because of the .NET framework, so maybe there’s no much the developers of this great tool can do.

I’ll start with the only missing option that doesn’t have to do with .NET, and it is that when selecting a given feed it has to show you always all of the posts of that feed, this causes a huge issue because of the .NET which I’ll explain later. I wish it to have an option that allow me to specify to not show any post when selecting a feed in the “Feed Subscriptions” pane, or to show me the last one or the one that was previously selected or perhaps the last 10, 20, 50, or whatever my preference is.

Now I’m going to talk about why the .NET framework is hurting this tool so much. First of all, I should say that I gave the opportunity to this tool to demonstrate that it can handle the massive amount of feeds that I need to track, and with the huge amount of postings that it gets everyday.

Right now I’m tracking 102 feeds on RSS Bandit and I may have something near the 81,000 or more entries in total. I know that at least a couple of feeds are closed, some have perhaps 1 post per month, others several per week, and some feeds (like the one for the IBM internal blogs) have several hundred per day.

Now you can imagine what do I need, and I have to say that .NET can’t handle it and looks like it will never be able to handle all of this. When selecting a feed, RSS Bandit gets unresponsive, I can’t do anything with it, and since the .NET comes from the Windows OS (I’ve XP installed) most of the times the task bar and everything is unresponsive. It doesn’t happen only when selecting a feed, it also happens when clicking on a menu option, a button, anything!!

When the above happens RSS Bandit consumes up to 366 MB of memory and the 99% of the CPU, something for about 2 or 3 seconds and often 7, 20, 30 seconds up to 3 minutes! And now that it is “idle” its memory usage is 356 MB !! Worst than Java I should say.

As you can see, .NET killed RSS Bandit from my point of view. It makes it require a machine with a processor of at least 50 GHz (the machine where I have it installed is a ThinkPad T41 with a Pentium M at 1.7 GHz and 1 GB of RAM) and at least 2 GB of RAM.

According to RSS Bandit’s page I have installed the latest version (1.3.0.42), this tells me that I won’t see any improvement in the tweak of .NET (if at all possible) in a long time. Perhaps the fastest way is to rewrite it in python or something like that?

Before I forget I’ll mention in here that RSS Bandit is using 73 MB to store the posts for all of the entries. I can’t say if that’s good or bad right now since I haven’t verified some of the other aggregators I’m reviewing.

Time to get rid of it !

This is something I wanted to do several months ago, basically because I really want to get rid of some of the RSS/ATOM aggregators (feed readers) from my machine. They give me a few problems but don’t wanted to just remove them right away without sharing with all of you what I think each one needs to improve on.

I will review the following tools which I have had installed for quite some time now:

• RSS Bandit
• SharpReader
• GreatNews
• Newzie
• RssReader
• Thunderbird

As you can see all of them are “standalone” applications and although I used Google Reader the first days it went live, I can’t say much about it because with all of the above tools I think they were enough for me.

I will post an entry per aggregator, and after that I’ll create a summary post with my conclusions. Time to start, lets get busy!

You already know that my name is César Miramontes and that I work for IBM, but you may not know what I do, that’s the reason of this post.

I joined IBM in 1998, just after I earned my Bachelor of Science degree in Computer Engineering.

I spent the first months as member of the team that was responsible for the Performance Tools for AS/400 product. Then I was the responsible of the Display Hardware Resources (DSPHDWRSC), Work with Hardware Resources (WRKHDWRSC), and Work with Hardware Products (WRKHDWPRD) CL commands, and the Hardware Resource APIs, the ones that start with QRZ, and a couple of years later I was also responsible of the ones that start with QGY.

In 2000 I joined the Domino for iSeries development team where I was for several years the “install guy”, and was responsible of the Domino for iSeries APIs and the CL commands to work with the servers.

I’m still part of the team, doing development and support for the product.

With that said, I just want to clarify that I’m not an IBM’s spokesman nor my weblog is a support page. Obviously I’ll be able to help or clarify any issue that anyone may have. And I think it’s worth mentioning that a lot of things for the platform and the product are already documented in one way or another.

That’s all for now, hope to see all of you visiting my weblog more frequently, I hope to have time to blog more frequently also.