Aug 29, 2007

Doing 'IT' for a bank -- that bad?

I came across the argument following Reddit link, here's a good quote to start:

... I am going to highlight the one thing that you should NOT do. I am hoping and praying that you are a lucky person and don't get sucked into the awful quagmire that I am about to outline below. ... If you are really, really unlucky, like most of us, you will leave school and 'go do IT for a bank'.

There's quite a bit of truth in much that Cron has written -- indeed, like in pretty much any other corporation that is not concerned with IT as a primary business area, there's little clarity of what and why to do this or that type of IT. Sure, just like in any larger organization, especially entrusted with a rather serious business of keeping other people's savings, technology choices tend to be rather "safe" ones.

Yet on the other side, there is an important aspect to banking that drives a lot of activity in some IT areas that are unlikely to be paid enough attention to in many other businesses.

Take, for instance, security, encryption, digital identity. Banks are the first organizations that have to be looking at that simply due to the nature of their businesses. So, if you're interested in *that* type of IT, and are not very much inclined in 'doing IT' for NSA or FSB (ex-KGB) -- banks are a much more peaceful place of employment that can still allow you doing enough research and technology development.

Or take the 24x7 resilience and uptime/on-line. Sure, you'd have to do the same for the military, medical, NASA. But a bank might be just another good place to try and do the same.

Or maybe mainframes is your thing. Maybe you were always fascinated with wonders of COBOL. Again, banks are the only few enclaves left where you can apply that skill and itch that urge.

So, to put it another way -- it is about what *you want to do*, not about how bad banks' IT is.

That said, there's a lot of mediocrity in banking IT, that is true. Things tend to be done by the book, in a very conservative way. So, if you know that you can't cope with any of that at all -- no, by all means, don't go do IT for the bank.

Aug 26, 2007

Numbers for the rest of us

Ever since Apple released Keynote rumors were circulating all over the fan sites as to when a full truly Mac office suit would be released. This was addressed bit-by-bit. First, Keynote was joined by Pages. Then both Keynote and Pages have received a small table management face-lift. Finally, this year Numbers has been released.

There's been many reviews of Numbers. Since I've been working with numbers most of my professional life, I felt the urge to give it a go and see how Numbers cuts numbers for me.

First great departure for many long-time Lotus 1-2-3 and MS Excel users is disappearance of work sheets. Sheets are still there, but they really refer to sections of your work book -- sort of like virtual pages in the file that may or may not fit a single A4 piece of paper.

At the same time you're getting the immense flexibility as to where and how you want to lay out your data. You would not notice this if all you need is a plain data listing (but then you may not be using the right tool), but if you, say, need to have details table and several summaries and maybe a chart or two to go along -- Numbers will really start shining for you.

Instead of providing you with a preset grid, you get just a page into which you can place as many individual tables as you like. This is the case of OLE finally done right, if you wish. And yet you can cross-link data in all these tables (since they act similar to worksheets of a traditional workbook).

This means that no longer do you need to play with column widths, merge cells or stretch your fields over expanses to achieve something like desired layout. And you get lots of samples of more or less typical spreadsheets to start using/learning/building.

Numbers also understands and interoperates with quite a few external sources/data formats. Excel is generally not a problem to import and integration with Address Book application lands a sweet spot for those who need to create simple lists/phone directories/etc.

That said, at this point it is very weak in the formulas support department covering only about 150 most frequently used ones. It lacks support for such "popular" data analysis tools like Excel Pivot Tables and filtering data is a bit of a bitch, since dynamic filters a-la Excel are also not available. Last but not least -- it is not blazing fast, especially once you start copying your formulas and/or data across the columns and rows.

Another bit of a problem and seemingly an oversight is the fact that you can't build a formula with keyboard only -- you *must* use a mouse or else key in cells and range names by hand.

It would seem that Apple could do much better following in the steps of KHTML to WebKit and back experience that allowed them to build a very good and standards-compliant browser quickly. I am pretty sure that both KSpread and Gnumeric could provide a great spreadsheet engine that already covers much more mathematical ground (especially Gnumeric).

Last note is on file format. When looked at on a console it is apparent that it is neither OOXML or ODF, but rather an XML dump of NSMutableArray Objective-C data. While this may have been done to speed the development and implementation, this will likely be a holding factor for going any further -- Apple would have to tackle on the interoperability issue with other spreadsheet vendors -- be it OOo, MSFT, KOffice or Gnumeric.

Aug 1, 2007

Взгляд: Россия нашла замену Microsoft

«Для нас важно, чтобы параллельно с устранением пиратства развивались программы с открытым кодом»
Давно пора. Хочется надеяться, только, что ето не попытка выторговать лучшую цену из Редмонда.