Tag Archives: software

Horrible Excel save times on USB

I don’t have the fastest USB drive, in fact, it’s probably crapish, primarily the horrible latencies it has.

But when saving a ~450KB .xls file takes 2 minutes, using Excel 2007, something can’t be right.  Copying the file to a local harddrive and saving there only takes one second.  Copying the file back, another second (more or less).  Evidently Excel is doing some crazy seeking whilst writing the file or similar.  But why is it so bad???  Surely, these days, it could easily write the entire file to memory before physically writing it to disk?

The joys of enterprise class systems

Enterprise class systems.  I prefer to call them unnecessarily ridiculously complicated, bloated, and just generally overly restrictive applications.

Well, I have to deal with Oracle for one of my courses.  It’s a wonderful database, I mean, the basic free edition only requires 2GB of disk space to install, and only consumes around 880MB of RAM whilst running, with no connections or databases (beyond what the installer puts there) or anything.  I can’t being to imagine the “true enterprise edition” and how much resources it needs.  Contrast it with MySQL, which has far more modest requirements of like 40MB install and uses under 30MB here.

I admit, I have not bothered trying to configure Oracle.  I don’t really want to, to be honest.  But one would imagine that the free non-commercial version is probably aimed at students or similar, and thus would be configured towards such setups (as opposed to a dedicated database server).

I haven’t used it much, but already, one of my beefs with it are auto-incrementing fields.  It’s a common thing to do, so crappy systems like MS Access include an easy to select AutoNumber type, and MySQL allows you to specify auto_increment in the CREATE TABLE statement.  Being an enterprise class DBMS, Oracle chooses to use a far superior method – you need to make a sequence, attach it to the table or whatever, and then use a trigger to update the sequence when a new record is inserted.  The fact that it requires more work means that the DBA can charge more hours for configuring an Oracle database, so it’s obviously better (except to the firm (aka suckers) paying for all this).

Though I guess Oracle at least has triggers, whereas MySQL doesn’t (or I haven’t seen them at least).

Oh, and if the above isn’t enough, their wonderful license agreements with free products are wonderful too.  I’m currently writing this because I’m waiting for some 20MB component to download over this 6KB/sec internet connection, because other applications couldn’t include this Oracle component.  They make it easy to grab the component too, well, I mean, it’s free, so I guess the people getting it can suffer right?  Need to register (or use BugMeNot) and download won’t go in download manager cause it needs cookie based authentication.  Urgh, how lovely.

I had someone ask me a bit about a query in Teradata (another enterprise class DBMS).  For MySQL, a simple SELECT GROUP_CONCAT(name) FROM people query would’ve done the job.  Obviously this enterprise class system doesn’t have that cheap GROUP_CONCAT function.  It does have a POSITION function for finding the first instance of something in a string, however, but it doesn’t have some sort of REVERSE_POSITION (InStrRev?) type function, for finding the last instance.  I guess if it had a string reverse function, it wouldn’t be so much of an issue, but it doesn’t even have that.  Hooray for the huge selection of available string functions there…

Thunderbird losing downloaded mail

Okay, this is irritating.  I just wanted to check some mail I’ve been expecting, but Thunderbird seems to have deleted all my stored emails for an account.  So it seems that its solution is to download absolutely every mail in the inbox again – yes, that’s >1000 emails it’s trying to download over my slow-ass connection.  Not to mention some mails where “newbie computer users” attach some 2MB Word document, a small image being most of it (at least Thunderbird allows you to restrict downloading message to a certain size).  And closing Thunderbird during this process causes it to restart every time it starts up again.  And I can’t seem to prioritise it to grab the latest mail or whatever.  WHY OH WHY??????!!!!

(above image shows my slow download speed)

I thought I selected to keep offline storage so it would use less bandwidth when loading emails.  Evidently, this means to refetch everything every now and then.  Maybe I should archive stuff…

Update: so I did go to move old emails to an archive folder, but it seems that Thunderbird is unable to figure out that I’m merely moving stuff, so when I actually try to access the folder, it goes to download everything again.  Well, as it’s not important, I guess I can leave it going whilst I’m doing something else later, but it’s a bit annoying.

I guess I really shouldn’t be complaining about free software – oh well.

Office 2010

Installed Office 2010 on all three of my computers yesterday (my main computer, parents’ computer and my netbook), after finding a way to activate Office without using the KMS service.  Bit of a shame that our MSDNAA doesn’t have Office, though they do have Project, Visio and Access (and maybe some other Office apps).

Initially installed 64-bit Office on my main machine, but later realised that MS doesn’t provide a 64-bit version of its common controls library (interestingly, it seems to be one of the few issues with 64-bit as opposed to 32-bit), so had to uninstall and install the 32-bit version.  Am a bit surprised with MS not including support for common controls in 64-bit Office – maybe someone will make some 64-bit components, which emulate the 32-bit behaviour (though, that has an issue that users of your “Office macro apps” will need to have this library installed).

Otherwise, I don’t terribly see that much of a difference between 2007 and 2010, so I guess upgrading is kinda moot (but of course, it’s cool, no?).  Heck, my workplace is still migrating from 2003 to 2007.

F***ing Web Installers

Why the f*** do some companies/people love web installers?  I mean, seriously, you download this 1MB installer, the only purpose of it being that it serves to download the real 21MB installer package.  Why can’t they just link to the installer package directly and eliminate this stupid intermediary step of downloading and running a web installer?

By the way, I wanted to download a new version of Skype to see if it resolved a weird issue I’ve been having, which is why I’m ranting.  It seems that Google loves to use web installers for Chrome too.

So why do I hate web installers?  Apart from seemingly offering nothing above traditional full package downloads (except maybe showing ads or similar) and the annoyance of having to download two things, it means you don’t actually have a copy of the installer locally.  That is, if you wish to reinstall a program, you need to download the damn thing again.  Or if you want to install it on a second computer, you can’t simply use an installer you’ve already downloaded (because the web installer usually deletes (or puts into a temp folder) the main installer package).

Also, our Uni has a fast internet connection (plus downloading from Google doesn’t count towards quota), so I like downloading Chrome from there, rather than through my slow home connection (and use up bandwidth).  Now, if you’ve got a web installer, this somehow becomes a little more difficult to do… (luckily Google actually does provide a standalone installer, though, conveniently, they don’t provide an easily findable link (well, I guess you can Google it)).

Now there are plenty of other reasons to hate web installers.  I’m sure there’s heaps more, but here’s a few:

  • Most don’t provide many options for selecting proxies; can also be an issue with some firewalls
  • More difficult, if not impossible, to use a download manager/accelerator
  • Some don’t even provide pausing functionality
  • Speed control can be more difficult (something a download manager like FDM could do)
  • Not really a direct fault of the web installer, but I find many mirrors actually mirror the web installer rather than the full installer, which means that if the server is down, or the link between you and the server is horribly slow, you’re kinda stuck

Now, I guess some applications, like Cygwin, it makes sense to have a web installer.  Having all Cygwin packages would be pretty big, and most people won’t even need anywhere near all of them.  In this case, it provides the benefit of reducing the amount of packages downloaded, though for apps like Chrome and Skype, this is stupid.
So please!  Please refrain from loving these stupid web installers so much!