Posts in category "software"

I wanted to export my Newsgator feeds into OPML and experiment with SearchFox, who kindly gave me an account, to take part in the beta program.

Only one slight problem, Newsgator doesn't support OPML export. Strange but true.

I am using the Newsgator Online RSS reader and simply want to order my most important, must-read blogs (i.e mine) at the top. It doesnt seem like an unreasonable requirement. From a cursory glance at the documentation, it is not immediately clear if I can even do this in Newsgator. This is a little irritating as I now have to do some work to scroll down to 'Oracle' or rename the folder as 'AAAOracle'. And, yes, you're right, I am very lazy.

Then I stumbled across a blog article on SearchFox which is another Web based RSS reader (like Newsgator) but with an interesting personalisation and recommendation engine which automatically bubbles your favourite, most frequently read articles and blogs to the top of your reading list.

This sounds interesting. It is a pity that SearchFox is another Web 2.0 application that is in beta and currently open to a selected few who must apply for an account. Still, I sent them a polite email citing my recent review and asked for access to evaluate their product. I even used a Gmail address to prove my Web 2.0 credentials. I sent the email yesterday and am now polling my Google Inbox every ten seconds.

I stumbled across SearchFox on a arbitrary blog search on feedster about RSS and it struck me that, in the past, I would use Google (Web, Groups), TuCows, SourceForge, Freshmeat to find OpenSource, free RSS readers.

However most of these sources are relatively out of date and had failed to unearth SearchFox when I was recently looking for RSS readers. The fast moving blogosphere is much better suited to finding these sort of hip, new Web 2.0 applications.

I am worried about those developers at WordPress. They have to eat food, drink coffee and wear clothes but how are they ever going to make any money ? There isnt even a Donate button anywhere on the site.

WordPress.com provide me with a hosted blogging platform which I think it is very good; better than Blogger, better than Bloglines, better than Yahoo 360', better than most of the competition.

The service provided by WordPress costs me absolutely nothing so represents excellent value for money and I would recommend the service to any of my blogging friends (if I had any).

WordPress have to provide servers, manage those machines, implement resilience, scalability and high availability, develop code, do boring things like backups, testing, fix bugs, worry about the business plan, buy laptops, S & M (sales & marketing, not the other one) and what do I give them in return ? A load of feedback (mainly negative) about minor, trivial things that don't work and waste their time whenever I have made an idiotic mistake (quite frequently).

Now, WordPress are planning to add extra features and functionality (customised CSS and templates, more themes and plug-ins, hosting on your server, statistics) which will cost money but they have also pledged that the current functionality will remain free. The sad fact is I am very unlikely to pay them anything for additional add-on features, ever.

This main reason is because I am quite happy with the existing product. Secondly, any additional contribution would have to be minimal as I could pay my ISP an extra 4 GBP per month to add PHP and then would be able to run WordPress.org with total control over everything. This would be more work for me but would probably be fun and an interesting experiment anyway.

Maybe I am not typical, maybe there are hundreds of frustrated WordPress bloggers out there with cheque books poised waiting for the two tier service to be announced. For the sake of the freeloaders like me, let's jolly well hope so.

Dont know where, dont know when but SonicStage 3.3 is available.

Includes ATRAC Lossless format, ripping WMA CDs, wider range of bitrates for ATRAC3Plus and MP3.

However, SonicStage remains the only software application in the modern world without a 'Check for updates' option.

I used to work for Ingres (in London) who were a fantastic company to work for. Amazingly, they are the only company I have ever worked for to use newsgroups for internal technical discussions and knowledge sharing instead of email aliases. I once read that processing an individual email costs a company 10 cents.

In the early 90's, Ingres was under commercial pressure from another large relational database vendor, Oracle. Instead of responding to this challenge, Ingres tended to 'fiddle while Rome burned', discuss the API naming convention by committee and stoutly defend the technical purity of page level locking (Oracle supported row level locking and capitalised heavily) from a lofty ivory tower.

Eventually, Computer Associates took over Ingres which most staff viewed as the end of the world. The truth was that CA saved Ingres from Chapter 11 (bankruptcy).

Although I had never visited Alameda and Islandia (Ingres' and CA's respective corporate headquarters), I still have this vision of hard nosed businessmen in sharp suits invading the Alameda campus to interview all these bearded techies in sandals.

On the day of the takeover, Oracle parked a truck outside Ingres' UK offices with a billboard 'Oracle are hiring now'. The story got a lot of coverage in the UK computer press. Once again, superb marketing from Oracle.

The Ingres engineers left the company in droves, formed self-help groups and arranged annual wakes to commemorate the anniversary of the black day.

CA subsequently rebranded the product OpenIngres but it largely disappeared from view into CA's vast portfolio of thousands of different software products.

So, it was nice to see Ingres back in the news this week as CA announced that Ingres Corporation will be once again be a separate company and the product will be available as an Open Source database.

Jeff Moss article about the commercial and free versions of Toad and the incredibly tenacious, persistent breed of salesperson bred by Quest Software got me thinking about the Oracle DBA tools I use.

• People
• SQL*Plus
• Statspack
• putty

People are important because people have developed the application, people are using the application, people are managing the servers, people are managing the database and intelligent people have configured that very expensive storage array.

These people know a lot about the application, the history of the project, the successes and failures, the lessons learned, the architecture, the infrastructure whereas I may know, quite literally, nothing about the same subjects.

I have seen people using Toad (and similar GUI based Oracle development tools) very effectively, multi-tasking, flipping between windows at breakneck speed. Sometimes it makes me quite tired just watching them.

However, I prefer SQL*Plus to do most of the work because :-

• SQL*Plus provides a complete report of my session, with timings, query plans, statistics. No need to frantically try to write it all down.
• The SQL script can be incrementally developed and is repeatable. It is quite easy in Toad to execute an arbitrary set of mouse-clicks so it is not clear which of the six 'alter session' statements were actually in effect when you finally got the optimal results.
• SQL*Plus is the 'vi' of the Oracle world. It is the one Oracle tool you are guaranteed to have access to, everywhere. The sys admin may not allow you to install commercial software without a valid license or may use another tool you have never seen before.
• SQL*Plus can use bind variables, provide the query plan (with or without retrieving the 6 million records) and provide the vital statistics (consistent gets) just the same as the GUI tools. Almost everything you need in fact.

Statspack is important because it is an Oracle package. Oracle will maintain and develop statspack for the latest features available in 10g. Statspack produces reports in a standard format which can easily be analysed by others (colleagues, DBA's, even Oracle). Statspack can also be configured to run automatically at regular intervals. If there is always a problem with the overnight ETL at 03.30, I would rather statspack gets the overtime and gathers the performance metrics rather than me sitting there in the middle of the night.

Statspack tells the truth, the whole truth and nothing but the truth. It is incredibly tempting to use an Oracle monitoring tool like OEM or Spotlight, obsessively watching the screen, drilling in on what appears to be an expensive query and your salvation. The problem is that you don't know the query in question is only spawned once a quarter for the accountant's financial report and runs in batch. The query could take 37 minutes and no-one would care. It is simply not important. It is not what the users are complaining about.

Worse, because your refresh interval is 5 seconds, you are blissfully unaware that you have missed the crucial SQL query with literal SQL that takes 1.7 seconds but is executed hundreds of times per second. Statspack does.

Setting the default statspack level to level 7 means that statspack can (retrospectively) produce the query plan for any problematic SQL statement identified in the summary report. This is handy where you may have long winded SQL statements where the summary report has tantalisingly truncated the SQL text just as the WHERE clause starts.

Putty is important so you can run O/S utilities (prstat, glance, topas, top, iostat, vmstat) to monitor the actual database server during the investigations. If the server is a development server hosting multiple Oracle instances with 2 CPU's running at 100% and saturated disks, then the performance of the application will be impacted, no matter what wizardry you weave.

Dear Bill

I live in London (near England) and would like to buy Microsoft Money and Microsoft Office. My preferred method of obtaining the goods would be to download these programs from your Web site and pay using a credit card.

A small discount to reflect the reduced administration costs, packaging and margin taken by the retailer would be nice but not essential.

However, when I attempt to buy these Microsoft products in the UK, I am redirected to third party Web sites (Amazon, Dabs, PC World etc) or I can delay the purchasing decision by downloading a 60 day trial version.

Buying the products from another outlet means that I have to do extra work just to make the order. This delay will be irritating and I might even consider using OpenOffice which I can download for free and start using now.

In fact, the only disadvantage of OpenOffice is that I anticipate my children handing in a piece of homework that Microsoft Word/PowerPoint/Excel is unable to convert correctly. Consequently their fantastic effort will receive a mark of zero and a detention as the teacher will not accept 'But, Sir, my Dad is an OpenSource evangelist' as a valid excuse.

Once I have ordered from Amazon, I then have to wait for the goods to be shipped, pray that Royal Mail doesn't mislay them and the postman doesn't leave the package unattended on my doorstep in the pouring rain (just because I didn't tip him last Xmas).

When I finally receive the goods, I then have to unwrap a large box containing fresh air to finally get my hands on the CD-ROM. There is no hardcopy user guide included as all the product documentation is now available online.

Note that Microsoft Money only costs 20GBP so comes supplied in a smaller box with less fresh air than Microsoft Office which costs 90GBP and, obviously, comes in a larger box with more fresh air.

I would prefer any response from Microsoft to be in the form of a old fashioned handwritten letter instead of this new fangled email technology.

Kind regards

Norman Brightside (Mr.)

When I started this blog, I simply composed the posts in the Blogger editor which was adequate. Until one day, when I lost the complete text of a draft posting due to finger trouble. As I laboriously re-typed my masterpiece, I wished I had a blog editor with the infinite undo, auto-save and all the other features of Emacs.

However, composing the drafts in the Blogger editor was useful as I could edit drafts from anywhere and then publish the blog very easily.

I then looked at Writely and Writeboard which fit the bill but are really intended for collaborative writing on the Web and don't have any integration with Blogger.

The Qumana Blog Editor also looked very interesting as it includes integration with Blogger and built-in support for Technorati tags but still was essentially a cut-down Word lookalike interface.

Then I realised I had the perfect blogging editor sitting right under my nose all the time - Emacs. I can use all of Emacs powerful text editing features and simply save the draft text on my Web server using ange-ftp.

Adding Technorati tags is easy using Marshall Kirkpatrick's BlogTags bookmarklet.

The only thing Emacs is missing is the ability to seamlessly publish to Blogger and another minor irritation is the fact that some whitespace gets jumbled when pasting the text into Blogger.

However, Emacs being Emacs, some kind person has created a Lisp package (weblogger.el) that provides integration with Blogger although I haven't actually tried it yet.

And please don't ask why I don't use the Blogger for Word extension. I can simply think of nothing worse. I wouldn't wish it on my worst enemies.

After briefly evaluating Furl, del.icio.us and Blinklist, I finally decided to ditch Furl and spurn the advances of del.ico.us in favour of Blinklist as my preferred one stop shop for all my social bookmarking needs.

Furl was my first experience of 'social bookmarking' but, as I became more comfortable with the idea of tagging Web pages, I found the Furl interface is simply horrible. There are simply too many key clicks required to add a single tag let alone two ! And if you want to add a brand new category, it takes so long that you have almost lost the will to tag the page and forgotten why it was of interest in the first place. It looks like the tagging functionality was added on as an afterthought.

I then experimented with del.ico.us as loads of people seem to rave about it and it appeared to be the longest established technology. After all, 50 million Web 2.0 aficionados and satisfied del.ico.us users can't be wrong, can they ?

However, just imagine that you are a first time user and you encounter the del.ico.us home page for the first time. The initial impression of the home page is absolutely terrible. What is it ? What does it do ? How do you use it ? How do you get help ? Where is the FAQ ? Or the tutorial ? Also, there is the confounded sub-domain - del.ico.us - and the oh so clever pun which I took three weeks to grasp.

After my previous post on social bookmarking, a nice gentleman from Blinklist spammed - sorry commented - on my blog to tell me about the existence of Blinklist so I felt obliged to evaluate Blinklist as well.

Blinklist's interface looks modern and clean, well designed and thought out as though they employ proper Web designers who actually use the system themselves ('eat your own dog food'). Contrast that with the initial page presented by del.ico.us which looks like an undergraduate knocked it up during a lunch hour and a manager said 'Looks good - let's run with it!'

It is true that I had some teething problems with Blinklist; for example when the tag editor kept insisting on SHOUTING AT ME and the odd tag got duplicated. However, I used the feedback form and tried to provide some constructive feedback to the Blinklist development team. Almost immediately, I found myself in direct email contact with a lead developer (well he could be the CEO for all I know) who passed on my suggestions and got them addressed very quickly. Sometimes they appeared to be making releases as quickly as I was providing feedback.

One of the best features of Blinklist is the tagging of pages is really quick and intuitive. Suitable tags are suggested automatically (which are usually adequate) and it is trivial to add as many new tags as you want - quickly. Blinklist also has a facility to import del.ico.us archives which worked fine for me when importing my massive archive of 15 pages. Your mileage may vary if you have 5,000 del.ico.us pages.

Although Blinklist doesn't keep a Web archive of the saved pages (like Furl) this doesn't really bother me as I can use the Internet archive if it really comes to it.

Of course, Blinklist has some disadvantages. I presume that Blinklist are a small startup and they are the new kid on the block in what is already a crowded marketplace. Worse, it is obvious that those clever people at Google labs will inevitably be entering this space (RSS Reader, My Search History) in the not too distant future.

Secondly, Blinklist has a relatively small user base (compared with Furl, del.ico.is, spurl et al) although this is compensated by the more intelligent, discerning type of people using the service and subsequent higher quality of the Blinklist content.

I used to use Yahoo Bookmarks which maintains a list of Web sites that I could access from any computer. This was a nice idea but I found I didnt use (or maintain) the bookmarks regularly and the links gradually fell into a state of disrepair. For my most frequently accessed Web sites, I would simply type the start of the address into the browser and simply let auto-complete do its work which was quicker.

A couple of weeks ago, Yahoo launched My Web which overhauled the bookmarks functionality and added the ability to save a copy of the Web page so that the referenced content was preserved in the event that the Web site was subsequently moved or deleted. However, My Web still uses a hierarchical system of folders to store the bookmarks. For example, the Manchester United home page would typically be stored in a folder named 'Sport-Football-Manchester United'.

Then I discovered social bookmarking (Furl, del.ico.us, spurl) which also saves the state of the Web page and maintains a personal, searchable archive (just like My Web). However, the key difference with Furl (et al) is that all Web pages for all users are stored on a central server. For any Web page, Furl can then quickly display similar, related, associated pages which have also been stored by different Furl users. This is what those Web 2.0 people call 'social bookmarking'.

Another difference is that Furl'ed pages does not use a hierarchical structure. Instead, stored pages are simply associated with various tags. So, in the earlier example, the Manchester United home page might be tagged as 'Sport', 'Football' and 'Manchester'. Note that, unlike My Web, the bookmarked Furl page can be associated with multiple categories. Tagging is the key to effective searching with Furl.

This feature is really useful. For example, I am currently interested in using Microsoft OneNote as a single repository to store information, emails, jottings, to do lists, articles, and even Outlook notes. OneNote has a couple of disadvantages; it uses a proprietary format and is a commercial product. So I searched SourceForge and Freshmeat for an OpenSource alternative without much success. Then I discovered KeyNote which looked promising but wasn't much different from using outline mode in Emacs.

Then I stumbled across EverNote and, using Furl, immediately got directed towards some more interesting Web based organiser applications like BackPack, JotSpot and possibly even TiddlyWiki all of which immediately get furled.

Another subtle way that Furl affected my usage was that previously I might stumble across a Web site of interest but could not be bothered to file it as (subconsciously) I didn't want to clutter up my nicely organised bookmarks. With Furl, I tended to add the Web site regardless and simply tag the page as 'Of interest'.

Another possible use for Furl is to capture ideas for blogs. When you see a Web site of interest, simply tag it as 'Ideas for Blog' together with a short comment.