• Andrew (unregistered) in reply to Anonymoose
    Anonymoose:
    Seriously, you posted a word document? That's a pretty big WTF right there.

    I would consult OpenOffice to see if can load the file. If that didn't work, I'd remove all non-ASCII characters from the file with Perl and read it.

  • (cs)

    So my DB knowledge is kind of lacking, which hasn't really helped my current job search. For #1 I would create a table that has 3 columns: ID - A unique ID for this node Parent - The ID for this node's parent Data - the actual data

    So if I want to find the children of node 1, SELECT Data,ID FROM TreeNodes WHERE Parent=1

    But that seems like it might be a pain to traverse the tree. Lots of searching. There a less WTFy way?

  • Zygo (unregistered) in reply to Andrew
    Andrew:
    Anonymoose:
    Seriously, you posted a word document? That's a pretty big WTF right there.

    I would consult OpenOffice to see if can load the file. If that didn't work, I'd remove all non-ASCII characters from the file with Perl and read it.

    I would do a quick Google search on the phrase "download full questions/response". I would then read through and filter out the search results to determine which are relevant and trustworthy to the problem at hand. I would also consult with my developers/DBA, colleagues, and internal knowledge databases to see if there’s a more specific best-practice solution.

  • Friggerock (unregistered)

    Of course the questions on the test are confusing. Duplicate conflicting and ambiguous statements are to be expected in a requirements document :P

  • Zygo (unregistered) in reply to Friggerock
    Friggerock:
    Of course the questions on the test are confusing. Duplicate conflicting and ambiguous statements are to be expected in a requirements document :P

    HR droid: "The candidate did not attempt to commit murder after reading the questions." PH boss: "Tell him he's hired!"

  • (cs) in reply to unklegwar
    unklegwar:
    The funny thing (to me) is that the readability of that first entry is on par with the readability of many TDWTF articles.

    Spelling? For the weak. Missing a verb or three? Inconsequential.

    Stones. Glass houses. Etc etc etc.

    You know what they say, if it spell checks, ship it!

  • Chris (unregistered)
    As it turned out, the Candidate’s experience didn’t much beyond English.

    And speaking of which...!

  • (cs) in reply to Davis
    Davis:
    Can whoever takes these screen shots, please turn off their damn ClearType (a.k.a. MSFuzzyType) settings!

    Damn thats a hard image to read! Just what I need to see when trying to read black text on a white background.. a bunch of rainbow-looks-like-an-anti-aliased-jpeg garbage image.

    I love the daily WTF, but that image is horrible, where are those damn goggles...

    Hm... spontaneous jab at Microsoft, complaints about having to read black text on a white background, general whineyness...

    ...guys, I think we've got a Linux Weenie!

  • (cs) in reply to Irrelevant
    Irrelevant:
    Davis:
    Can whoever takes these screen shots, please turn off their damn ClearType (a.k.a. MSFuzzyType) settings!

    Damn thats a hard image to read! Just what I need to see when trying to read black text on a white background.. a bunch of rainbow-looks-like-an-anti-aliased-jpeg garbage image.

    I love the daily WTF, but that image is horrible, where are those damn goggles...

    Hm... spontaneous jab at Microsoft, complaints about having to read black text on a white background, general whineyness...

    ...guys, I think we've got a Linux Weenie!

    Improbable; most of the modern antialiasing schemes for X11 look fuzzier even than the impressively fuzzy ClearType; fuzzy is in.

  • drew (unregistered)

    did he get the job?

  • Lee (unregistered)

    Hey, I live in Cleveland! Oh wait, I live in the UK one :(

  • OutOfToiletPaper (unregistered)

    Sorry guys, which way to the nearest supermarket?

  • StickyWidget (unregistered) in reply to Zygo

    ATTENTION ALEX!!!

    OK, I'll admit that Question 3 is a little confusing. It's either a REAL simple question, using the solution proposed in post#147417 by Zygo above, or its a trick question involving object oriented design and implementation. I'm pretty damn sure it's a trick question.

    The problem is that object.equals(..) only returns true if the object EXACTLY equals the object in the HashSet (i.e. a.equals(b) iff a IS b). In other words, they have the same memory address, it's a dumbed down pointer comparison. So, simply passing an object with the same values as the original will not work, they are simply different, and will evaluate as such.

    Since this is an object oriented question, I would override the "equals" method of the object, and replace it with something like:

    ...
return (this.intMember == someIntParameter  &&
this.stringMember.equals(someStringParameter);
..

    This would allow any test for equality throughout the entire application to be properly assessed by ANY operation that required a true/false equality test. Game-Set-Match.

    The real WTF is that it took 63+ Comments by people who are supposed to be SME's to finally point this out. Oh, and this guy is a doofus for not actually solving the problems. My professors in college would have written "You did not solve the problem, you receive no credit, and you should really consider a career in management."

  • Isospace (unregistered) in reply to StickyWidget

    You also need to define a hashCode method

  • Faux News (unregistered) in reply to Anonymoose
    Anonymoose:
    Seriously, you posted a word document? That's a pretty big WTF right there.
    Yeah, where the hell's my PDF?
  • Stefan W. (unregistered)

    An alternative SQL-approach to q2:

    SELECT course, MIN (inst) as i1, ' ' as co from inst GROUP BY course HAVING COUNT (*)= 1 UNION
SELECT course, MIN (inst) as i1, MAX (inst) as co from inst  GROUP BY course HAVING COUNT (*)= 2 UNION
SELECT course, 'comitee', '' as co from inst GROUP BY course HAVING COUNT (*)> 2 ORDER BY course;
  • Jeremy Weiskotten (unregistered)

    Here's how you solve #3 (with a JUnit test for a "program").

    I'm still amazed at how many Java developers simply don't understand equals/hashCode. Read Effective Java already!

    import java.util.HashSet;
import java.util.Set;

import junit.framework.TestCase;

public class HashSetExampleTest extends TestCase {
	
	private final class Dude {
		private final int age;
		private final String name;
		
		public Dude(int age, String name) {
			// TODO: Validate that age is > 0
			// TODO: Validate that name is not null or empty
			this.age = age;
			this.name = name;
		}
		
		public boolean equals(Object obj) {
			if (!(obj instanceof Dude))
				return false;
			
			Dude otherDude = (Dude)obj;
			// two dudes are equal if they have the same name and age
			return otherDude.age == this.age && 
				otherDude.name.equals(this.name);
		}
		
		public int hashCode() {
			int hash = 1;
			hash = hash * 31 + this.age;
			hash = hash * 31 + this.name.hashCode();
			return hash;
		}		
	}
	
	private final Set<Dude> dudes = new HashSet<Dude>();
	
	@Override
	protected void setUp() throws Exception {
		super.setUp();
		dudes.add(new Dude(15, "Joe"));
		dudes.add(new Dude(99, "Fred"));
	}


	public void testContains() {
		assertTrue(dudes.contains(new Dude(15, "Joe")));
		assertTrue(dudes.contains(new Dude(99, "Fred")));
	}
	
	public void testDoesNotContain() {
		assertFalse(dudes.contains(new Dude(15, "Fred")));
		assertFalse(dudes.contains(new Dude(99, "Joe")));
	}
}
  • Casey (unregistered)

    I had an interview I was doing on the telephone, where after the second technical question the guy stopped us and said "look, these questions are a bit pointless, why don't you tell me what your project is about, and I'll tell you how to implement it" :S

  • Watson (unregistered)

    Honestly, the first time I looked at that email shown there I parsed it as saying that they were on a retarded search...

  • Jon W (unregistered) in reply to StickyWidget

    Dude, you failed. If you override equals(), you have to also override hashCode(). As I pointed out about 20 comments back, btw. And someone else before me.

  • Alan (unregistered)

    I wrote a technical test for a c# developer position at my company. It is a simple paper & pen test we give applicants on thier first interview. It only contains questions that shouldn't need google to answer.

    One candidate had a CV which really impressed us, and had everything we were looking for. They scored 1 out of 9 on thier test. For example:

    Q9: Write a C# function that returns true if the input is 0 or greater than 5.

    Applicants answer:

    int a; if(a = 0 or >5) { return(true); } else { return(false); }

  • Jerome (unregistered)
    Alex:
    As it turned out, the Candidate’s experience didn’t much beyond English.

    On the other hand, Alex's experience didn't even that far.

  • (cs) in reply to vt_mruhlin
    vt_mruhlin:
    So my DB knowledge is kind of lacking, which hasn't really helped my current job search. For #1 I would create a table that has 3 columns: ID - A unique ID for this node Parent - The ID for this node's parent Data - the actual data

    So if I want to find the children of node 1, SELECT Data,ID FROM TreeNodes WHERE Parent=1

    But that seems like it might be a pain to traverse the tree. Lots of searching. There a less WTFy way?

    This is the 'normal' way of doing things, and generally acceptable. A less intuitive but more efficient method is to allocate 'left' and 'right' numbers to each node in the tree, starting at the top node, and working round in a circle (check the link and you'll see what I mean). This then makes your queries truly trivial, e.g. to find all children of a node, or all leaf nodes. However, it does involve renumbering some items when you add a new node.

    There's a better explanation out there somewhere, but I can't find a better link than this one; I'm a database expert, not a Google expert.

    http://www.sitepoint.com/article/hierarchical-data-database/2

  • (cs) in reply to Shinobu
    Shinobu:
    Davis:
    Damn thats a hard image to read!
    a) You're grossly exaggerating. b) Turning off ClearType is work and actually makes it look worse on most displays.

    Cleartype has to be set up for the kind of monitor you use (RGB or BGR TFT), if an image is captured with one cleartype mode and displayed on the opposite kind of monitor it is INCREDIBLY hard to focus on.

    Someone with a CRT monitor should never enable cleartype (use standard font smoothing instead), as it relies on the way TFTs arrange the colours in their pixels. It is theoretically possible to make a version of cleartype for CRTs, but it hasn't yet been done.

  • anon (unregistered) in reply to Sharkie
    Sharkie:
    With the general quality of coders in the employment pool, I'd be absolutely thrilled to find one that would actually perform web searches to locate solutions, as well as work with internal resources towards the best solution.
    Agree. But they aren't looking for a coder, they're looking for an architect. An architect is someone who's been around the block, and actually knows stuff- not just a code monkey who decides he's been doing it for a while and deserves a better title. Or worse, a code monkey who's not very good at it, and thinks he must be good at something, and "architecting" is it. (I think our applicant is the second type.) OK, I take that back. Most "architects" are just code monkeys who want a better title. Hence, this site exists.
  • (cs) in reply to Thief^
    Someone with a CRT monitor should never enable cleartype (use standard font smoothing instead),

    I would, but some idiot decided it should be disabled for normal font sizes with no option to enable it. "standard font smoothing" is disabled between 7 and 13 points, i.e. between 9 and 18 pixels, on windows. Other systems exhibit similar behavior. (Actually, I think it's per-font, since Consolas is smoothed at all sizes, the sizes I listed above were for Arial)

    Cleartype also uses different (better) hinting rules than standard font smoothing. It would be nice if there was a "cleartype gray" (i.e. enable the nicer hinting and smoothing at all sizes, without any subpixel wizardry) setting. Since there's not, cleartype still looks better on CRTs than 'standard', even with the colors.

  • anon (unregistered) in reply to Mario
    Mario:
    The answer for Q1 should contain a note about the number of parents. You could assume one parent per child, but it may be more than one. Remember, reader, you have 2 parents!

    No one is asking about your family tree. In computer science, a tree structure is a directed acyclic graph with indegree 0 or 1.

  • dkf (unregistered) in reply to Thief^
    Thief^:
    It is theoretically possible to make a version of cleartype for CRTs, but it hasn't yet been done.
    It's much more difficult to do, since CRTs tend to distribute their color cells in two dimensions, whereas LCDs put them in a line (so only 1-D error propagation is needed in that case). Given that CRTs are basically on the way out, the effort required to make a cleartype-alike for them is a waste.
  • Corey (unregistered) in reply to zip
    zip:
    Did anyone here actually read that word document? The questions (espcially #3) are horribly written. If this wasn't because of anonymization (and I don't see why this would need to be anonymized) then I'd say the real WTF is him writing 2 pages of BS when he should've told them to shove it as soon as he saw how badly their standard "test" was written.

    I dunno... if you weed out all the jobs that have clueless or annyoing HR, I imagine you'll soon get most of your nutrition by raiding discarded pizza boxes for cheese.

  • Edward Royce (unregistered)

    Hmmmmm.

    I keep thinking I'll take the time out to get into Java again.

    Is there a decent GUI toolset for Java out there or does it still all suck?

    Captcha "bathe". Uhhh. Darn it! How does it know??

    :)

  • Look at me! I'm on the internets! (unregistered) in reply to StickyWidget
    StickyWidget:

    The real WTF is that it took 63+ Comments by people who are supposed to be SME's to finally point this out.

    What are you snarking at? Of course you have to override equals. This goes without saying, and is standard practice when creating classes.

    Besides, your code is wrong. You're not overriding equals, you're overloading it.

    The signature for equals is

    public boolean equals(Object other)

    not

    public boolean equals(int someIntParameter, String someStringParameter)

    the proper code would be

    public boolean equals(Object other)
{ 
  //  wrong type, not equal (could throw exception instead)
  if (! (other instanceof ThisClass)) return false; 
  
  // cast to appropriate type.
  ThisClass _other = (ThisClass)other;
  
  
  return (this.intField == _other.intField) && this.stringField.equals(_other.stringField);
}

    You might be able to eliminate the casting if you use generics, but I'd have to test it out first.

  • Michael (unregistered) in reply to Zygo
    Zygo:
    SELECT CourseNo, 
    CASE WHEN cnt <= 2 THEN i1 ELSE 'Comittee' END AS Instructor1,
    CASE WHEN cnt  = 2 THEN i2 ELSE NULL       END AS Instructor2
FROM (
    SELECT CourseNo, MIN(Instructor) AS i1, MAX(Instructor) AS i2, COUNT(*) as cnt
        FROM "Course Instructors" GROUP BY CourseNo
) AS q ORDER BY CourseNo;

    If there are 43 gazillion instructors teaching a single course, then maybe the COUNT(*) is too slow and should be replaced by a subquery that uses EXISTS and/or whatever extension syntax this particular database uses to implement offsets and limits. But as for how to do it, I don't know; I would have to look it up.

    This should work for MySQL:

    SELECT CourseNo, 
IF(COUNT(Instructor) <= 2, MIN(Instructor), 'Comittee') AS Instructor1, 
IF(COUNT(Instructor) = 2, MAX(Instructor), NULL) AS Instructor2
FROM Course_Instructors
GROUP BY CourseNo ORDER BY CourseNo

    Assuming that the Instructor field would never be a non-NULL empty string.

  • Michael (unregistered) in reply to Edward Royce
    Edward Royce:
    I keep thinking I'll take the time out to get into Java again.

    Is there a decent GUI toolset for Java out there or does it still all suck?

    That depends, by "decent" do you mean small and simple, or powerful and flexible?

    I've been using Swing for years, and it has always been decent, and recently has become much better in terms of native look and feel. Two new specs expected in Java 7 (Application framework and Bean Binding) will further simplify things for developers.

  • Edward Royce (unregistered)

    Hmmm.

    Mostly powerful and flexible.

    Last time I did anything with Java, loooooong ago, it was a hassle trying to do UI in AWT and Swing. Has it matured since then?

    Thanks in advance for any info. Right now I'm mostly shifted to C#/ASP.NET but I've been looking at JBoss and really like what I see so I was thinking about hedging my options and making sure I proficient in a non-Microsoft dev package.

  • Rich (unregistered)

    People, CASE statements are basic SQL and if you don't know how to use them, you're probably not really using SQL, you're just reading and writing data.

    Rich

  • Rich (unregistered) in reply to vt_mruhlin
    vt_mruhlin:
    But that seems like it might be a pain to traverse the tree. Lots of searching. There a less WTFy way?

    No. There are a few cheaty ways and shortcuts but they tend to be more WTFey. With SQL, it really helps if you can limit the depth of the tree.

    Rich

  • (cs) in reply to Rich
    Rich:
    People, CASE statements are basic SQL and if you don't know how to use them, you're probably not really using SQL, you're just reading and writing data.

    Rich

    I don't know about you, but that's what I'm using my database for. I'm not writing SQL for the pure circle-jerkery of it.

  • Rich (unregistered) in reply to bobday
    bobday:
    Rich:
    People, CASE statements are basic SQL and if you don't know how to use them, you're probably not really using SQL, you're just reading and writing data.

    Rich

    I don't know about you, but that's what I'm using my database for. I'm not writing SQL for the pure circle-jerkery of it.

    Personally, I like to make the most of the relational aspects and the ability of the query optimizers to pre-process certain types of data more efficiently than can be done on the front end code.

    There is nothing wrong in using a database just to read and write data but if that's all you're doing with it, you're not really using SQL.

    Rich

  • (cs)

    I can understand why a lot of people don't see how the Architect candidate's answer would be considered so bad. I don't mean this as a knock against anyone but you do have to have a good amount of systems design experience to get it. Allow me to spell it out. This person:

    a) Treated certain complex concepts with no agreed-upon best practices - such as trees in an RDBMS - as trivial code-monkey tasks that a few minutes with Google could solve.

    b) Wasted more time, effort, and words explaining how he'd solve the problems and/or why he didn't need to solve them, than he would have spent actually solving them (assuming he knew what he was doing).

    c) Apparently didn't know SQL. This is completely OK for a programmer, but it's a big no-no for an architect. SQL is ubiquitous, it exists in almost every environment and is a critical part of any business system. DBAs may help to design individual tables, optimize indexes, set up maintenance plans, that sort of thing - but they are not responsible for designing the entire data model top-to-bottom. That is an architect's job.

    d) Criticized the last question and threw his hands up in the air instead of either making reasonable assumptions or asking for clarification. Yes, I know the question was worded badly. Guess what - so are most requirements and functional specs. Resolving the ambiguities and inconsistencies in a fuzzy spec are primarily the job of an architect. They probably didn't intentionally word the question poorly, but even still, this person's reaction reveals terrible business instincts.

    Bottom line: judging by the answers, I can only assume that he's spent the past "8.5+ years" actively avoiding any real work or learning. On the job, he probably gets things done - eventually - but maybe not as well as someone else would, and certainly not without a whole lot of attitude. Pass.

  • (cs) in reply to Rich
    Rich:
    No. There are a few cheaty ways and shortcuts but they tend to be more WTFey. With SQL, it really helps if you can limit the depth of the tree.
    String sets are reliable and fast, although if you consider triggers to be WTFey (I don't) then I guess the implementation is WTFey. And they do involve a certain depth limit, but the limit is arbitrary and set by the design; one of our databases supports 50 nesting levels (never seen more than 10 used) but you could design a mapping to support hundreds if necessary. Just avoid the recursive trigger and use an iterative one instead.

    SQL Server 2005 also gives you CTEs with recursive queries, which is a far more elegant solution, although of course it's going to be platform-specific.

    Those are certainly not the only ways, just the most well known... most of the ideas aren't cheating or WTFey either, they're covered in off-the-shelf books on data structures and algorithms. They aren't as simple and pretty as the generic parent-child join, but no production system is ever as simple as it's freshman academic equivalent.

  • Anonymous (unregistered) in reply to Davis
    Davis:
    Can whoever takes these screen shots, please turn off their damn ClearType (a.k.a. MSFuzzyType) settings!
    You and your CRT screen can go to hell!
  • Nelle (unregistered) in reply to Zygo
    Zygo:
    Admittedly my Java-Fu is all grasshopperish, but I don't really get question #3 either.

    AFAICT there's no method on HashSet named contains that accepts an Object and returns an Object from the HashSet. There are a number of methods inherited from various places with similar names to contains (contains, containsKey, containsValue...) which all return bool, and there's a method named get that does take an Object and return a similar Object from the hash table (using Object's hashCode method, presumably).

    If you just want to know if the object is in the HashSet, that should be trivial: just construct an object identical to the one that was inserted, and fire it into contains(). Someone, somewhere, in the implementation must "locate the original Object" in order to determine whether contains() returns true.
    [...]

    I believe the goal was for the applicants to write a custom equals( Object o ) method for the custom object that was to be stored in the hashmap ...

    But I have to agree with the guy, no "contains" method ever returns an Object ...

  • (cs) in reply to Aaron
    Aaron:
    c) Apparently didn't know SQL. This is completely OK for a programmer
    Since when? I've worked with SQL pretty much daily ever since I started getting paid to write code. Maybe, if you're writing video games...
  • (cs) in reply to Zygo
    Zygo:
    "Create a query which will print the all the courses"

    When reading this I get visions of reams of paper coming out, as the entire course (including all lecture notes, textbooks, answer guides...) comes flying out of one of those high-volume campus-sized laser printers, all because someone wanted "all the courses" printed.

    I also notice the question doesn't specify what happens when zero instructors are teaching a course. Inevitably there will be many attempts to implement answers here. It will be interesting to see how many cope with that case.

    My first one doesn't--guess I need a left join after all:

    SELECT CourseNo, 
    CASE WHEN cnt <= 2 THEN i1 ELSE 'Comittee' END AS Instructor1,
    CASE WHEN cnt  = 2 THEN i2 ELSE NULL       END AS Instructor2
FROM (
    SELECT CourseNo, MIN(Instructor) AS i1, MAX(Instructor) AS i2, COUNT(*) as cnt
        FROM "Course Instructors" LEFT JOIN "Courses" USING (CourseNo) GROUP BY CourseNo
) AS q ORDER BY CourseNo;

    You are the first one that gets the whole question right :)

    The other person that also answered made a similar query, but he didn't notice that the question asked for all courses to be displayed, and most probably got misled by the fact there the example table had no course with no instructors on it.

    The actual question lacked a word or two (probably lost while being reworded), but I think it's still quite clear.

    "Create a query which will print the all the courses and if one instructor is present display their name [their name == the instructor's name]"

  • (cs) in reply to StickyWidget
    StickyWidget:
    ATTENTION ALEX!!!

    OK, I'll admit that Question 3 is a little confusing. It's either a REAL simple question, using the solution proposed in post#147417 by Zygo above, or its a trick question involving object oriented design and implementation. I'm pretty damn sure it's a trick question.

    The problem is that object.equals(..) only returns true if the object EXACTLY equals the object in the HashSet (i.e. a.equals(b) iff a IS b). In other words, they have the same memory address, it's a dumbed down pointer comparison. So, simply passing an object with the same values as the original will not work, they are simply different, and will evaluate as such.

    Since this is an object oriented question, I would override the "equals" method of the object, and replace it with something like:

    ...
return (this.intMember == someIntParameter  &&
this.stringMember.equals(someStringParameter);
..

    This would allow any test for equality throughout the entire application to be properly assessed by ANY operation that required a true/false equality test. Game-Set-Match.

    The real WTF is that it took 63+ Comments by people who are supposed to be SME's to finally point this out. Oh, and this guy is a doofus for not actually solving the problems. My professors in college would have written "You did not solve the problem, you receive no credit, and you should really consider a career in management."

    IsoSpace:
    You also need to define a hashCode method

    Yeah, the HashSet onject will internally create a hashcode of the object passed by parameter, and then it will compare this hashcode to the hashcodes of all objects on the set.

    Hell, that's the whole point of using a HashWhatever instead of a List: not having to use the equals() method on all objects on the list.

    And you are right about contains() returning false if it's not the same onject on the same memory location. That's why the "correct" answer is overloading the equals() function so it will return true if the objects are identical, even if they are not on the same memory location :P

    Addendum (2007-07-31 15:00):

    P.D.: damn, and, of course, it would be nice to remember to overload HashCode() too, but the HashCode contract does not require it ^^

    http://java.sun.com/j2se/1.4.2/docs/api/java/lang/Object.html#hashCode()

    "It is not required that if two objects are unequal according to the equals(java.lang.Object) method, then calling the hashCode method on each of the two objects must produce distinct integer results. However, the programmer should be aware that producing distinct integer results for unequal objects may improve the performance of hashtables."

    Addendum (2007-07-31 15:03): That damned 5 minutes edit limit makes me make mistakes -.-

    You do need to overload HashCode too, specially on the circumstances describe by this question.....

  • Jasper Mantooth (unregistered)

    Kind of off topic but:

    Why do people constantly bitch about the quality of this site? It takes less effort to hit ctrl-w than it does to click 'Add Comment' and spray your verbal diarrhea. That annoys me more than the 'First!' and 'Captcha' posts...

    PS: captcha: alarm -- wake up guyz! har har har! lolz!1

  • (cs)

    The headhunting part of the article reminds me of a dice posting I saw a few years back. They were recruiting for a hardware position, and one of the requirements was familiarity with "scuzzy controllers."

  • (cs)

    Oh boy! I can't wait to hire this architect to create high level designs and have no f'ing clue what any of it means!

    Do you want an architect who can BS and give uninformed design decisions... or do you want someone who can demonstrate knowledge of the underlying components he is drawing on the whiteboard?

    My favorite part is his comment about "Hey if you want me to solve this problem I'll just consult my internal knowledge database." No you won't, or you would have already done it.

  • (cs) in reply to Rich
    Rich:
    Personally, I like to make the most of the relational aspects and the ability of the query optimizers to pre-process certain types of data more efficiently than can be done on the front end code.

    There is nothing wrong in using a database just to read and write data but if that's all you're doing with it, you're not really using SQL.

    Rich

    Alright, since you actually made a serious reply, I'll clarify a bit:

    I think I use the relational aspects of SQL, since I have as many left outer joins as the next guy, but I've never found much use for the case statement.

    I don't have any 1-or-2-to-1 relationships (what a headache), so I typically just use the extra query required to get all the data in a many-to-one relationship. Also, I can't do too much fancy processing in database queries since it has a tendency to interfere with localization.

  • (cs) in reply to rycamor
    rycamor:
    What does our IT Architect consider low-level?
    At a guess, anything remotely concerned with "implementation". He sounds like the kind of person who believes that people who actually get their hands dirty with real code can't possibly expand their focus out to see the big picture - but are nonetheless always the ones to blame when the design turns out to be a big pile of manure.

    In fact, he sounds rather like this guy...

Leave a comment on “You Get What You Ask For”

Log In or post as a guest

Replying to comment #:

Log Out

« Return to Article